diff --git a/AGENTS.md b/AGENTS.md index c6c8d6a..42633b6 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -81,8 +81,8 @@ Use these terms as the default implementation contract for Video Workbench slide | --- | --- | --- | | `子会话` / `session` | A user-visible separate Codex session for brief / orchestration work, started from a clean prompt rather than inherited orchestration history. | `codex_app.create_thread` to create, `codex_app.send_message_to_thread` to continue, `codex_app.read_thread` to inspect. | | `继续前子会话` | Continue the already recorded visible child session instead of creating a new one. | Send the next instruction to the recorded `threadId`. | -| `Agent` / `SubAgent` | An internal one-shot worker owned by the child session; it should not appear as a separate user-visible Codex session. | `multi_agent_v1.spawn_agent(agent_type="worker")`; one worker handles exactly one slide. | -| `无 review` | Review is disabled for the named scope. | Do not write `sNN-vN-review.md`; record only packet / return / ledger / task-log evidence. | +| `Agent` / `SubAgent` | An internal one-shot worker used for test/audit, exception handling, or explicitly isolated slide work. It is not the default production-light carrier. | `multi_agent_v1.spawn_agent(agent_type="worker")`; one worker handles exactly one slide when used. | +| `无 review` / `review: none` | Review is disabled for the named scope. This is the production default after the test / audit line has established the project's core look. | Do not write `sNN-vN-review.md`; record only compact status, image paths, and task-log / handoff notes needed for continuation. | Do not use `codex_app.fork_thread` as the default carrier for production child sessions. Forking copies completed orchestration history from the parent thread and can contaminate the child session's context. Use `fork_thread` only when the user explicitly wants an inherited-context branch. @@ -91,21 +91,32 @@ If the default tools are not already exposed in a new session, use `tool_search` Default preflight for a slide/session run: - confirm the target visible session id or create exactly one new clean visible session with `create_thread`; +- when creating a visible child session, set or request a user-facing title beginning with `子会话：` followed by the project / package name, so the session is distinguishable in the Codex UI; - scan the target slide folder for the next unused `sNN-vN-*` version; - set `Overwrite allowed: no`; -- record whether review is disabled or enabled; -- require the child session to launch internal one-shot SubAgents for slide work. +- read the project phase from `projects//project.md`; new projects default to `test/audit` until the user explicitly authorizes formal production; +- record the active line, either the project default or a user-authorized override for a named session / scope; +- record review mode; production-light defaults to `review: none`; +- use direct-child production in `production-light`; launch internal one-shot SubAgents only in `test/audit`, explicit review, repair, or exception scopes. + +Use two operating lines: + +- `test/audit`: use when building or debugging the workflow, proving a new project look, testing core shots, or investigating failures. Local review, richer ledgers, and longer handoffs are allowed. +- `production-light`: use after core shots have reached the expected look. This is the default normal production line once authorized in `project.md`. It keeps the content line (`brief`, `prompt`, generated image, `slides/slides.md`) and compact recovery receipts, but omits routine audit reports. Production orchestration defaults: - The main session dispatches the approved scope, records the parent thread id as the callback target, then stops routine waiting. It should not poll the child thread for routine completion. -- The child session writes `child-session-handoff.md` first, then sends a concise callback to the parent with `codex_app.send_message_to_thread`. -- The callback payload must include status, completion gate, resolved versions, handoff path, image paths if complete, and any problems or user decisions needed. -- After the callback wakes the parent, the parent reads the child session's short handoff and continues. It should not audit prompt / return / ledger / image details when the child session reports `complete` with `completion_gate: pass`. +- The production-light child session directly writes briefs, prompts, images, `slides/slides.md` updates, and a compact handoff. It does not launch Agent/SubAgent workers by default. +- The child session completes all allowed project writes, `slides/slides.md` updates, narrow required checks, and compact handoff writing before callback. +- The child session sends a concise callback to the parent with `codex_app.send_message_to_thread` as its final action. After a successful callback, it must not write or verify additional project files. +- The callback payload must include status, topology, resolved versions, handoff path, image paths if complete, and any problems or user decisions needed. +- After the callback wakes the parent, the parent reads the child session's short handoff and continues. It should not audit prompt / image details when the child session reports `complete` with no problems. - The old sleep / poll schedule is a fallback only when callback tooling is unavailable or the user explicitly requests polling for a named run. If used, record the fallback in the task plan. -- The child session must run the mandatory completeness gate before returning: append-only version, no overwrite, expected files present, review present or absent as instructed, one owning worker per slide/version, references loaded or blocked, and compact handoff written. -- Instruction / visual review is optional. Run it only when the task asks for review; `无 review` means no `sNN-vN-review.md`. -- A slide `Agent` / `SubAgent` is one-shot. It receives one packet, handles one slide/version, writes one return, then becomes terminal / closed. Do not continue or reuse a closed Agent; repairs, retries, and next slides require a new Agent. +- The production-light child session keeps only the mandatory write guard: resolve the next unused version by scanning existing files, do not overwrite, record image paths, update required status files, perform only narrow required checks, and write a compact handoff before callback. Do not run a routine final verification pass. +- In `test/audit`, repair, or exception scopes, the child session may run the fuller completion gate: append-only version, no overwrite, expected files present, review present or absent as instructed, one owning worker per slide/version when workers are used, references loaded or blocked, and compact handoff written. +- Instruction / visual review is optional. In production-light, the default is `review: none`: generated images are recorded as production-pass outputs without per-image user review, and the user may request an append-only retry later if the result is not usable. +- A slide `Agent` / `SubAgent`, when used, is one-shot. It receives one packet, handles one slide/version, writes one return, then becomes terminal / closed. Do not continue or reuse a closed Agent; repairs, retries, and next slides require a new Agent. Re-investigate runtime mechanics only when a tool is unavailable or errors, when an Agent becomes a user-visible session, when duplicate workers target the same slide/version, when target files collide, or when the user explicitly asks to revise the mechanism. @@ -113,14 +124,17 @@ Rules: - Use `projects//tasks//` for local orchestration records. - The main session talks with the user and defines worker profile, running mode, slide scope, and review expectations. -- Child sessions write per-slide briefs and may handle at most 3 slides by default. -- A slide agent handles exactly 1 slide per invocation. +- Production-light child sessions write per-slide briefs, prompts, images, and compact handoff directly; they may handle at most 8 ordinary production slides by default. Reduce to 1-3 slides for identity-heavy, repair-heavy, core-risk, or context-expensive scopes. +- Test/audit child sessions may launch slide agents; a slide agent handles exactly 1 slide per invocation. - The default worker profile for current MV / video work is `video-slide-production`. - Reserved future profile `ppt-page-production` may use `docs/cases/`, but must be designed when the first real PPT task needs it. +- For cover pages, PPT pages, science-video pages, title cards, explainer pages, or any slide that needs later human layout / typography / post-production, add a separate `sNN-vN-design.md` before `brief`. `design` is for the user and records visual concept, layout, text content, typography/color guidance, post-production notes, constraints, and references. `brief` remains the execution card for Codex / image generation and must not absorb human-facing design or post-production notes. +- `sNN-vN-design.md` is Chinese by default. If an English design companion is needed, write `sNN-vN-design-en.md` beside it; the English file supplements but does not replace the default Chinese design. +- For design-gated slides, write `design` first and wait for user confirmation before writing `brief`, `prompt`, or generating the image, unless the user explicitly authorizes the full chain in one task. - `visual-system/` is the production authority for slide work. - Style continuity must use `visual-system/style/style-prompt-master.md`. - Character continuity must use `visual-system/characters/character-reference-registry.md` and active character image references. - Visible character identity cannot be generated from text-only prompts when identity matters. - Slide prompts must use the `gpt-image-2` Skill Advisor / host-native workflow. - New slide-task execution must append the next unused `sNN-vN-*` version and must not overwrite existing slide artifacts unless the user explicitly authorizes an overwrite repair for named files. -- Review defaults to local-only. Do not call Minimax MCP or other external vision/review services unless the user explicitly authorizes that task. +- Review defaults to `none` in production-light and local-only in test/audit or explicit review tasks. Do not call Minimax MCP or other external vision/review services unless the user explicitly authorizes that task. diff --git a/VIDEO_WORKBENCH.md b/VIDEO_WORKBENCH.md index 27716e5..5746ef0 100644 --- a/VIDEO_WORKBENCH.md +++ b/VIDEO_WORKBENCH.md @@ -4,7 +4,7 @@ This project is the local execution workspace for dimensional video and slide-style output. -It turns accepted GPT V2 planning Markdown and user-provided sources into project notes, local execution plans, visual-system assets, slide/page execution folders, generated assets, and handoff material. It is optimized for practical continuation, not heavy governance. +It turns accepted GPT V2 planning Markdown and user-provided sources into project notes, local execution plans, visual-system assets, slide/page execution folders, generated assets, and compact handoff material. It is optimized for practical continuation and production throughput, not heavy governance. ## 2. Collaboration Roles @@ -52,6 +52,8 @@ If a series needs more structure, create `projects/_series/` when the first real `projects//project.md` is a lightweight Codex-maintained note. It should record: - project status; +- lifecycle phase from `docs/workflows/project-lifecycle.md`; +- workflow phase: new projects default to `test/audit`; switch to `production-light` only after explicit user authorization; - medium type; - source and reference links; - which `intake/` files are GPT V2 planning Markdown, source notes, reference notes, or background notes; @@ -69,9 +71,16 @@ Visual projects must have: projects//execution-plan.md ``` +Global lifecycle rules live in: + +```text +docs/workflows/project-lifecycle.md +``` + This is a living Codex-maintained execution control file. It records: - the current medium branch; +- the current lifecycle phase when it affects execution; - the current execution focus; - which visual-system assets need materialization; - the current small-batch iteration strategy; @@ -189,17 +198,25 @@ slides/ s01/ s01-metadata.json s01-unit-spec.md + s01-v1-design.md # optional; required for cover/PPT/science-video/title/post-production pages + s01-v1-design-en.md # optional English companion; default design remains Chinese s01-visual-brief.md s01-v1-prompt.md s01-v1-image.png - s01-v1-review.md + s01-v1-review.md # optional; test/audit or explicit review only s01-v2-prompt.md s01-v2-image.png - s01-v2-review.md + s01-v2-review.md # optional; test/audit or explicit review only ``` All files inside `sNN/` must carry the `sNN` prefix. Iteration files use `sNN-vN-type.ext`, so files from the same version sort together. +`design` is a human-facing design specification, not an image-generation execution card. Use `sNN-vN-design.md` for cover pages, PPT pages, science-video pages, title cards, explainers, and any slide that needs later human layout, typography, text placement, or post-production. It records the design concept, intended text, placement, typography/color guidance, constraints, references, and post-production notes. `brief` stays compact and execution-facing for Codex / image generation. + +`sNN-vN-design.md` is Chinese by default. If an English design companion is needed, write `sNN-vN-design-en.md` beside it. The English companion is optional and never replaces the Chinese default. + +For design-gated slides, write `design` first and wait for user confirmation before writing `brief`, `prompt`, or generating images, unless the user explicitly authorizes the whole chain. Ordinary MV shot production may skip `design` and continue directly with `brief -> prompt -> image`. + New task execution must append a new slide version instead of overwriting an existing one. Resolve the next unused `vN` from existing `sNN-v*-*` files in the slide folder, and treat incomplete or superseded earlier versions as historical evidence rather than paths to reuse. ### Slide Task Orchestration @@ -219,20 +236,38 @@ projects//tasks// Task records are for orchestration only. Durable slide artifacts remain under `slides/sNN/`. +For production work that may span multiple child sessions, keep a compact continuation file: + +```text +projects//tasks//production-run-status.md +``` + +This file records the target scope, completed units, active child session, next cursor, latest callback, and stop conditions. It is the default recovery surface after parent-session compression. It is not a review report and should not duplicate prompts or image evidence. + Roles: - Main session: receives user instructions, defines worker profile / running mode / slide scope, starts child sessions, records the parent callback target, and reads handoffs after callback. -- Child session: reads shared context, writes per-slide briefs, starts one slide agent per slide, writes a child-session handoff, and callbacks the parent after the handoff exists. -- Slide agent: handles exactly one slide, writes the prompt, generates only when allowed, writes review only when required, and returns compact status. +- Child session: reads shared context and handles the assigned package. In `production-light`, it directly writes briefs, prompts, generated images, `slides/slides.md` updates, and a compact handoff before callback. +- Slide agent: optional worker for `test/audit`, explicit review, repair, exception handling, or intentionally isolated slide work. It handles exactly one slide when used. + +Operating lines: + +- `test/audit`: early project setup, core-shot proving, mechanism debugging, or failure investigation. Richer handoffs, ledgers, local review files, Agent workers, and parent inspection are allowed. +- `production-light`: normal production after core shots have reached the expected look and the user authorizes the switch in `project.md`. Keep the content line (`brief`, `prompt`, generated image, `slides/slides.md`) and compact recovery handoff; omit routine review, Agent workers, packet/return files, and audit records. Defaults: - Worker profile for current video / MV work: `video-slide-production`. -- Child session capacity: at most 3 slides. -- Slide agent capacity: exactly 1 slide. +- Design-gated slides: cover, PPT, science-video, title, explainer, and post-production-heavy pages require `sNN-vN-design.md` before `brief`. +- Child session capacity: at most 8 ordinary production slides; reduce to 1-3 for identity-heavy, repair-heavy, core-risk, or context-expensive scopes. +- Production-light topology: main session -> child session directly. +- Test/audit topology: main session -> child session -> optional slide Agent. +- Slide agent capacity when used: exactly 1 slide. - Execution is sequential by default; do not run child sessions or slide agents in parallel unless the user explicitly changes the rule. -- Parent orchestration is callback-only by default: after dispatch, the main session stops routine waiting; the child session wakes the parent with status, completion gate, resolved versions, handoff path, and image paths if complete. +- Parent orchestration is callback-only by default: after dispatch, the main session stops routine waiting; the child session wakes the parent with status, topology, resolved versions, handoff path, and image paths if complete. +- Callback is the child session's final action: all allowed project writes, `slides/slides.md` updates, narrow required checks, and the compact handoff must be complete before callback; after a successful callback the child must not write or verify additional project files. - Polling is a fallback only when callback tooling is unavailable or the user explicitly requests polling for a named run. +- Production-light review mode: `none`. - Future `ppt-page-production` work is reserved and may use `docs/cases/`, but it must not inherit video-slide rules blindly. The user controls task running mode: @@ -240,7 +275,9 @@ The user controls task running mode: - `prompt only`: create briefs and prompts; do not generate images. - `prompt and generate`: create briefs and prompts, then generate images for the task's recorded slide scope. -Task-level generation instructions count as user approval for the named scope and mode. Core, repair, identity-heavy, or otherwise high-risk slides may still pause if the task plan says they require sampling or explicit review. +Task-level generation instructions count as user approval for the named scope and mode. In production-light, generated images are recorded as production-pass outputs without per-image user review or routine final verification; the user can request an append-only retry later if a result is not usable. Core, repair, identity-heavy, or otherwise high-risk slides may still pause if the task plan says they require sampling or explicit review. + +For unattended production, the parent session repeats the callback loop from `docs/workflows/project-lifecycle.md`: update `production-run-status.md`, dispatch one child session, wait for callback, reload the minimal project/task state, then dispatch the next package if no stop condition is present. Exact parent context/token usage is not required for continuation because Codex may auto-compress the parent session; file-backed state is the authority. Visual-system rules: @@ -252,9 +289,10 @@ Visual-system rules: Review rules: -- Review defaults to local-only. +- Production-light review defaults to `none`. +- Test/audit review defaults to local-only unless the task says otherwise. - Do not call Minimax MCP or other external vision/review services unless the user explicitly authorizes that task. -- Non-core generated slides can be tracked in `tasks//task-log.md` and `slides/slides.md` instead of a full per-slide review. +- Generated production slides can be tracked in `tasks//task-log.md` and `slides/slides.md` instead of a full per-slide review. ## 9. Investigations @@ -306,7 +344,7 @@ gpt-image-2 Skill Advisor prompt/spec → load the real reference image into the conversation context → call host image_gen with the loaded image assigned an explicit role → copy the selected output into the project or investigation -→ record prompt, input reference path, output path, and review +→ record prompt, input reference path, output path, and compact handoff/status ``` This route was validated in: @@ -324,7 +362,7 @@ For every reference-capable generation: - the prompt must name the loaded image role, such as `Image 1: identity reference`, `Image 1: style reference`, or `Image 1: edit target`; - merely writing a local path in the prompt is not enough; - the output must be copied from the host-generated default location into the project, investigation, or other durable workspace path; -- the review must record whether the generation used the verified host reference-by-visible-context path. +- the compact handoff or task status must record whether the generation used the verified host reference-by-visible-context path; a slide-agent return receipt is required only when the task actually uses a slide Agent in test/audit, repair, exception, or explicit review mode. Reference selection defaults: diff --git a/docs/workflows/project-lifecycle.md b/docs/workflows/project-lifecycle.md new file mode 100644 index 0000000..612c53a --- /dev/null +++ b/docs/workflows/project-lifecycle.md @@ -0,0 +1,104 @@ +# Project Lifecycle + +## Status + +- Status: active global workflow rule. +- Scope: all `video-workbench` projects, regardless of whether the final artifact is an MV, explainer video, slide-style presentation, or another visual/video package. + +This document defines the common lifecycle. Project-specific files decide what each phase means for that medium. + +## Lifecycle Phases + +Use these phases in `projects//project.md`: + +| Phase | Meaning | Common exit condition | +| --- | --- | --- | +| `intake-cleanup` | Accepted GPT / user source material is repaired, normalized, and stored under `intake/`. | Source files are readable, named, and mapped to local execution roles. | +| `system-definition` | Source material is localized into project production systems. | Local control files and production authorities exist. | +| `pilot-validation` | Core units, risky visual/audio rules, and workflow mechanics are tested. | The user accepts the core route and authorizes production. | +| `production` | Units are produced in batches through the approved route. | All target units are complete, accepted or explicitly deferred, and no active production package remains. | +| `assembly-closeout` | Produced assets are assembled, exported, reviewed for final use, and archived. | Delivery/output artifacts and handoff records are complete. | + +The lifecycle phase is not the same as the operating line. A project in `production` can still temporarily use `test/audit` for a repair or failure investigation. A normal production package should use `production-light` after the route is approved. + +## Medium-Specific Outputs + +The phase names stay stable; the concrete work differs by medium. + +| Medium | `system-definition` examples | `pilot-validation` examples | `production` examples | +| --- | --- | --- | --- | +| MV / shot-image package | `visual-system/`, style master, character registry, `slides/slides.md`. | Core shots, character consistency, reference-loading route, callback route. | Sequential slide image packages. | +| Explainer video | Script structure, terminology, visual explanation rules, voice style, asset taxonomy. | One short script section, key diagram, sample narration audio, timing check. | Script sections, diagrams, narration audio, supporting slides. | +| Slide-style presentation | Page system, component patterns, visual hierarchy, case/source map. | Representative pages and transition/page-density checks. | Page batches and export-ready assets. | + +Do not copy GPT planning into duplicate project truth. The accepted `intake/` files remain sources; Codex-owned local execution rules live in `execution-plan.md`, `visual-system/`, `slides/`, `tasks/`, and medium-specific control files. + +## Production Run State + +For any production request that may continue across multiple child sessions, create or update a lightweight run state file: + +```text +projects//tasks//production-run-status.md +``` + +This file is the compression-safe continuation surface for unattended production. It should be short and current, not an audit report. + +It records: + +- project path; +- lifecycle phase; +- operating line; +- running mode; +- review mode; +- package size or validation override; +- total target scope; +- completed units; +- active child session and callback target; +- next unresolved cursor; +- stop conditions; +- latest callback summary. + +The parent session may be automatically compressed. After each child callback, the parent should reload the minimal source of truth before dispatching the next package: + +```text +AGENTS.md +VIDEO_WORKBENCH.md +docs/workflows/project-lifecycle.md +docs/workflows/slide-task-orchestration.md +docs/workflows/slide-task-templates.md +projects//project.md +projects//execution-plan.md +projects//slides/slides.md +projects//tasks//production-run-status.md +projects//tasks//task-log.md +``` + +Only read detailed prompts, images, old session records, or prior handoffs when the current handoff reports `blocked`, `failed`, `warning`, collision, ambiguous scope/status, or the user explicitly asks for inspection. + +## Unattended Production Loop + +When the user authorizes a full production run: + +1. Resolve the target set from the project fact table. +2. Write `production-run-status.md` and `task-log.md`. +3. Dispatch exactly one clean child session for the next package. +4. Stop routine parent polling and wait for callback. +5. The child completes all artifact writes, `slides/slides.md` updates, compact closeout checks, and `child-session-handoff.md` before callback. +6. The callback is the child's final action. The child must not write or verify additional project files after a successful callback. +7. On callback, the parent reads the child handoff and updates the run state. +8. If status is clean and target units remain, dispatch the next package. +9. Stop only when the target set is complete or a stop condition is reached. + +The parent does not need exact token/context usage to continue. It relies on file-backed reload after callback. A separate successor orchestration session is a fallback, not the default, and should be used only when the current parent cannot continue cleanly. + +## Stop Conditions + +Stop and ask the user when: + +- a requested scope cannot be mapped cleanly; +- a package would include an already accepted unit unless explicitly requested; +- a target output path already exists and overwrite was not authorized; +- required references or production authority files cannot be loaded; +- child callback reports `blocked`, `failed`, or `warning`; +- user confirmation is explicitly required for the validation package result; +- the next action would change global package-size defaults or lifecycle rules. diff --git a/docs/workflows/slide-task-orchestration.md b/docs/workflows/slide-task-orchestration.md index 19da9de..aed5aaf 100644 --- a/docs/workflows/slide-task-orchestration.md +++ b/docs/workflows/slide-task-orchestration.md @@ -14,7 +14,8 @@ In this workflow, `slide` is the execution unit. A slide may be a video frame un - Keep style and character continuity anchored in `visual-system/`. - Keep user review focused on real decisions, not every generated image. - Keep task evidence light but sufficient for continuation after context compression. -- Keep routine completeness and optional instruction review inside child sessions, so the main orchestration session can stay lightweight in production. +- Keep mandatory write guards inside child sessions, so the main orchestration session can stay lightweight in production. +- Separate test / audit work from normal production so routine production does not pay test-level token cost. ## Worker Profiles @@ -39,7 +40,7 @@ gpt-image-2 Advisor prompt/spec -> load the real reference image into visible conversation context when references are required -> call host built-in image generation with the loaded image named as Image 1 / Image 2 -> copy the host output into slides/sNN/ --> record prompt/spec evidence, loaded-reference evidence, host output path, and review/return evidence +-> record prompt/spec evidence, loaded-reference evidence, host output path, and compact handoff/status ``` Garden/API use is expected to be `no` unless the user explicitly changes the project policy. @@ -58,6 +59,32 @@ Every task plan must record: Worker Profile: video-slide-production ``` +## Operating Lines + +Use two lines: + +```text +test/audit +production-light +``` + +`test/audit` is for early project setup, first visual-system proving, core-shot trials, workflow debugging, failure investigation, or explicit user-requested review. It may use richer handoffs, local review files, Agent workers, packets, returns, ledgers, parent-side inspection, completion gates, and longer narrative logs. + +`production-light` is for normal production after core shots have reached the expected look. It is the default production line. It keeps only: + +- content artifacts: `sNN-vN-brief.md`, `sNN-vN-prompt.md`, generated image when allowed, and `slides/slides.md`; +- compact recovery records: child-session handoff and one-line task log entries. + +In production-light, review mode defaults to: + +```text +review: none +``` + +Generated images are recorded as production-pass outputs. They do not need per-image user review before the workflow continues. If the user later finds a result unusable, run an append-only repair or retry as a new version. + +Project phase is recorded in `projects//project.md`. New projects default to `test/audit`. Once the user explicitly authorizes formal production, switch the project phase to `production-light`; future runs then load the production-light workflow by default. A named session or slide scope can still override back to `test/audit` for local diagnosis. + ## Directory Model Use project-local `tasks/` for orchestration records: @@ -68,9 +95,11 @@ projects// / task-plan.md task-log.md + production-run-status.md # for multi-package production or compression-safe continuation session-01/ - child-session-plan.md child-session-handoff.md + # production-light default: no packets/returns + # test/audit or exception scopes may add: packets/ s05-v1-slide-agent-packet.md returns/ @@ -80,7 +109,7 @@ projects// s05-v1-brief.md s05-v1-prompt.md s05-v1-image.png - s05-v1-review.md + s05-v1-review.md # test/audit or explicit review only ``` Rules: @@ -88,6 +117,7 @@ Rules: - `tasks//` records orchestration and agent communication. - `slides/sNN/` remains the durable home for slide artifacts. - Task records link to slide artifacts by path and should not duplicate full prompts. +- `production-run-status.md` is the compact continuation surface for multi-package production. It records scope, completed units, active child session, next cursor, latest callback, and stop conditions. - Task records do not own the slide version number. Before each slide run, resolve the next unused `sNN-vN-*` version from the existing files in `slides/sNN/`, then write only that new version. - Existing slide artifacts are historical evidence even when incomplete or superseded. Do not overwrite them to make a later task look cleaner. - Suggested task id pattern: `YYYY-MM-DD-`. @@ -115,28 +145,33 @@ Responsibilities: - receive the user instruction; - define the worker profile, running mode, slide scope, and review expectations; -- create or update `task-plan.md`; +- read the project phase from `project.md`, unless the user explicitly overrides it for a named session/scope; +- create or update `task-plan.md` only with the active scope and dispatch/callback status needed for continuation; - split work into sequential child sessions; -- start child sessions one at a time; +- start child sessions one at a time and name them with a user-facing `子会话：...` title when the tool surface supports thread naming; - record the parent callback target for the child session; - read the child session's short completion handoff after callback; - ask the user only for blockers, core decisions, or requested sampling; - update `task-log.md` with compact status. +- update `production-run-status.md` after each callback when the task may continue across multiple child sessions. -The main session should not write per-slide briefs, final prompts, images, or routine reviews by default. In production mode it should also not redo routine child-session completeness checks unless the child session reports a problem. +The main session should not write per-slide briefs, final prompts, images, or routine reviews by default. In production-light it should also not redo routine child-session checks unless the child session reports a problem. Production orchestration default: - Launch or continue exactly one child session for the approved scope. - Record the parent thread id and require callback-only completion unless the task plan explicitly declares polling fallback. - After dispatch, stop routine waiting. Do not inspect the child thread for routine completion. +- In `production-light`, instruct the child session to work directly. Do not start Agent/SubAgent workers, do not write packet files, do not write return receipt files, and do not run a final verification pass. - The child session must write `child-session-handoff.md` before sending any callback. -- The child session sends a concise callback to the parent after the handoff exists. -- The callback payload must include status, completion gate, resolved versions, handoff path, image paths if complete, and any problems or user decisions needed. -- If a package would exceed 3 slides, split it into smaller packages unless the task plan explicitly defines a custom package size. -- Consume only the handoff's status, completion gate, review mode/result, agent lifecycle, and parent summary by default. -- Continue to the next approved slide/scope when status is `complete` and the completion gate is `pass`. -- Open detailed prompt, return, ledger, or image evidence only when the handoff reports `blocked`, `failed`, `warning`, missing completion-gate evidence, artifact collision, duplicate worker, user-requested audit, or explicit sampling/review. +- The child session must complete all allowed production-light writes, `slides/slides.md` updates, compact handoff writing, and any narrow file-existence checks before callback. +- The child session sends a concise callback to the parent only as its final action. After a successful callback, it must not write or verify additional project files. +- The callback payload must include status, topology, resolved versions, handoff path, image paths if complete, and any problems or user decisions needed. +- If a package would exceed 8 ordinary production slides, split it into smaller packages unless the task plan explicitly defines a custom package size. +- Consume only the handoff's status, topology, review mode, resolved versions, image paths, and needs-user field by default. +- Continue to the next approved slide/scope when status is `complete` and the handoff reports no problems. +- For unattended production, reload the minimal project/task state after callback and continue dispatching packages until the target scope is complete or a stop condition is present. +- Open detailed prompt, return, ledger, or image evidence only when the handoff reports `blocked`, `failed`, `warning`, artifact collision, duplicate worker, user-requested audit, or explicit sampling/review. Polling fallback is allowed only when callback tooling is unavailable or the user explicitly requests polling for a named run. If fallback polling is used, the task plan must record the reason, the status-check schedule, and the timeout / blocker policy before dispatch. @@ -146,31 +181,34 @@ The child session is the small-batch context holder. Default capacity: -- maximum 3 slides per child session; -- fewer slides for identity-heavy, repair-heavy, or context-expensive work; +- maximum 8 ordinary production slides per child session; +- fewer slides for identity-heavy, repair-heavy, core-risk, or context-expensive work; - sequential only, not parallel. This maximum is a default child-session chunk size, not a project-wide production cap. After the workflow is proven stable, the main session may dispatch additional child sessions or a broader user-approved scope while preserving the per-child-session capacity rule unless the task plan explicitly changes it. +The 8-slide default is based on accepted production-light compression validation in `projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/`: eight images accepted into the production pass by user rule, 174K tokens, 67% context, and 38m37s elapsed. The callback-final contract was verified in the child thread tail. The earlier 5-slide validation remains historical capacity evidence and exposed the callback-order defect, so child prompts must still state that callback is the final action. + Responsibilities: - read project rules and the task plan; - read shared project context for assigned slides; - read active `visual-system/` files needed for assigned slides; - resolve the next unused slide version and write one new `sNN-vN-brief.md` per assigned slide; -- create one slide-agent packet per slide; -- start exactly one slide agent per slide, sequentially; -- verify expected files were written; -- run the required completion gate before returning; -- run instruction review only when the task plan or packet requires review; +- in `production-light`, write `sNN-vN-prompt.md`, generate `sNN-vN-image.png`, and update `slides/slides.md` directly; +- in `production-light`, do not create slide-agent packets, do not start Agent/SubAgent workers, do not write return receipt files, and do not run a routine final verification pass; +- in `production-light`, finish all project writes and narrow required checks before callback; do not continue writing after callback succeeds; +- in `test/audit`, repair, or exception scopes, create one slide-agent packet per slide and start exactly one slide agent per slide, sequentially; +- in `test/audit`, repair, or exception scopes, verify expected files and run the required completion gate before returning; +- run instruction review only when the task plan or packet requires review; production-light defaults to no review; - write `child-session-handoff.md` with a compact parent-facing summary; - send the parent callback only after `child-session-handoff.md` exists. -If the runtime does not let the child session start a slide agent directly, the child session must record this as a runtime fallback requirement in `child-session-handoff.md`. The main session may then start the slide agent as a fallback, but that run must be reported as a fallback orchestration test, not as a full child-session-owned slide-agent test. +If a test/audit run needs an Agent and the runtime does not expose an Agent tool inside the child session, record this as a runtime fallback requirement in `child-session-handoff.md`. The main session may then start the slide agent as a fallback, but that run must be reported as a fallback orchestration test, not as a full child-session-owned slide-agent test. ### Slide Agent -The slide agent is the single-slide worker. +The slide agent is the optional single-slide worker for `test/audit`, repair, exception handling, explicit review, or intentionally isolated slide work. It is not the default production-light carrier. Default capacity: @@ -186,7 +224,7 @@ Responsibilities: - write the new-version `sNN-vN-prompt.md`; - generate an image only when the packet allows generation; - write `sNN-vN-review.md` only when required; -- write a compact return file. +- write a compact return receipt. Lifecycle rule: @@ -218,16 +256,16 @@ Flow: ```text child session writes brief --> slide agent writes prompt +-> child session writes prompt -> if visible character identity matters, load required character reference image(s) --> slide agent calls host image generation through the visible-context reference path --> slide agent writes review if required --> slide agent return +-> child session calls host image generation through the visible-context reference path +-> child session updates slides/slides.md +-> child session writes compact handoff and callbacks parent ``` For this workflow, "load required character reference image(s)" means the actual local image is made visible to the generation context before host image generation. A local filesystem path written in a prompt is traceability only and does not count as loading the image. -The slide agent must not treat "Garden/API not used" as a failure. The expected evidence is that the Advisor prompt/spec was prepared and the host image-generation path was used with the required visible-context reference. +When a slide agent is used, it must not treat "Garden/API not used" as a failure. The expected evidence is that the Advisor prompt/spec was prepared and the host image-generation path was used with the required visible-context reference. ### Prompt Only @@ -237,8 +275,8 @@ Flow: ```text child session writes brief --> slide agent writes prompt --> slide agent return +-> child session writes prompt +-> child session writes compact handoff and callbacks parent ``` Generation can happen later through a separate task. @@ -250,11 +288,12 @@ The user gives the main session concrete instructions for: - running mode; - slide scope; - generation permission; +- operating line; - review or sampling expectations. If the task says `prompt only`, no image generation is allowed. -If the task says `prompt and generate`, the task plan must record the authorized slide scope and exceptions. That task-level instruction is the generation approval for the named scope. Core, repair, identity-heavy, or otherwise high-risk slides may still require prompt-only handling or an explicit pause when the task plan says so. +If the task says `prompt and generate`, the task plan must record the authorized slide scope and exceptions. That task-level instruction is the generation approval for the named scope. In production-light, generated images are accepted into the production pass without per-image user review. Core, repair, identity-heavy, or otherwise high-risk slides may still require prompt-only handling or an explicit pause when the task plan says so. ## Visual-System Requirements @@ -285,55 +324,72 @@ docs/workflows/slide-task-templates.md Required per-slide artifacts: ```text +sNN-vN-design.md # required for cover/PPT/science-video/title/post-production-heavy slides +sNN-vN-design-en.md # optional English companion; default design remains Chinese sNN-vN-brief.md sNN-vN-prompt.md sNN-vN-image.png # only when generation is allowed and succeeds -sNN-vN-review.md # only when required +sNN-vN-review.md # only for test/audit or explicit review ``` +`design` is the human-facing page or cover design spec. It is for user review and later layout / typography / post-production. It may include visual concept, intended text, text placement, font/color guidance, case references, design constraints, and post-production notes. It is not the prompt and should not be treated as a runnable image-generation artifact. + +`sNN-vN-design.md` is Chinese by default. Use `sNN-vN-design-en.md` only as an optional English companion when needed. + `brief` is the universal execution card. It is written by the child session. `prompt` is an attachment to the brief. It should be runnable or easy to copy into another platform for cross-validation. -`review` is optional by default. Create it only for core trials, identity-heavy slides, repair targets, explicit user requests, uncertain failures, or durable accepted references. +`review` is absent by default in production-light. Create it only for test/audit, core trials, identity-heavy sampling, repair targets, explicit user requests, uncertain failures, or durable reference acceptance. -When review is required or suppressed, the task plan or slide-agent packet should say so explicitly. Character-consistency tests and other identity-sensitive generated trials default to local review unless the task plan disables it. +When review is required or suppressed, the task plan or slide-agent packet should say so explicitly. Production-light tasks default to `review: none`; test/audit tasks default to local review unless the task plan disables it. -## Completion Gate +Write `design` before `brief` for cover pages, PPT pages, science-video pages, title cards, explainers, or any slide that needs human text/layout/post-production decisions. Pause for user confirmation after `design` unless the task explicitly authorizes continuing through `brief`, `prompt`, and generation. Ordinary MV shot slides can skip `design` and continue directly through `brief -> prompt -> image`. -The completion gate is mandatory for every production or workflow-test child session. It is not the same as instruction review. +## Production-Light Closeout -The child session must confirm before returning: +Production-light uses a minimal closeout, not a full audit gate. + +The child session must: - assigned slide versions were resolved by scanning existing `slides/sNN/sNN-v*-*` files; - no target file was overwritten unless an explicit overwrite repair authorized that exact path; -- each assigned slide has the expected brief, packet, prompt, image when generation is allowed, and return files; +- each generated slide has a project image path recorded in `slides/slides.md` and the compact handoff; +- review files are absent when review mode is `none`; +- the child session handoff exists before callback. + +Do not run a routine final verification pass in production-light. Do not compute hashes, build file inventories, inspect generated images, or write return receipts unless the task explicitly asks for that evidence or a problem occurs. + +If any production-light closeout item fails, return `blocked`, `failed`, or `warning` and state what the main session should inspect. The main session then leaves lightweight production mode for that item and performs targeted inspection only. + +## Test / Audit Completion Gate + +Use the fuller completion gate only in `test/audit`, repair, explicit review, or exception scopes. The child session should confirm: + +- append-only version resolution and no overwrite; +- expected brief, packet, prompt, image when generation is allowed, and return files; - review files are present when review was required and absent when review was disabled; -- each slide has exactly one owning slide agent for that slide/version; +- exactly one owning slide agent for that slide/version when Agents are used; - each slide agent return is recorded and the agent lifecycle is terminal / closed; - required visible-context image references were loaded or the run blocked before generation; -- generated project images have recorded source paths and hashes when available; +- generated project images have recorded project paths and source paths; hashes are optional unless requested; - the child session handoff contains a compact parent summary. -If any completion-gate item fails, the child session must return `blocked`, `failed`, or `warning` and state what the main session should inspect. The main session then leaves lightweight production mode for that item and performs targeted inspection only. - ## Review Rule -Default review is local-only. +Production-light default review is `none`. Test/audit default review is local-only. - Do not call Minimax MCP by default. - Do not transmit private project images to external services unless the user explicitly authorizes it for that task. -- Completeness checking is mandatory and belongs in the completion gate. +- Completeness checking is a test/audit concern. Production-light uses only minimal closeout. - Instruction / visual review is optional and depends on the task plan. -- If review mode is `none`, do not write `sNN-vN-review.md`; the return and handoff must still state that review was intentionally suppressed. -- For character-visible slides, check whether the required character image reference was actually loaded. -- For style-sensitive slides, check whether the prompt used the active style module from the style master. -- For host-native generation, check whether the review or return records: +- If review mode is `none`, do not write `sNN-vN-review.md`; the handoff, and the return when a slide Agent was used, must still state that review was intentionally suppressed. +- For character-visible slides, the prompt or handoff should state the required character image reference role. +- For style-sensitive slides, the prompt should use the active style module from the style master. +- For host-native generation, production-light records only what is useful for continuation: - `gpt-image-2` Advisor prompt/spec evidence; - loaded visible-context image role such as `Image 1`; - - host generated-image source path; - - copied project image path; - - whether Garden/API was expected (`no` by default). + - copied project image path. Routine non-core generated slides can be tracked in: diff --git a/docs/workflows/slide-task-templates.md b/docs/workflows/slide-task-templates.md index 96c029b..c120391 100644 --- a/docs/workflows/slide-task-templates.md +++ b/docs/workflows/slide-task-templates.md @@ -5,10 +5,39 @@ - Status: active template reference. - Companion workflow: `docs/workflows/slide-task-orchestration.md`. -Copy these structures into project-local task and slide files as needed. Keep them compact; do not turn them into full audit reports. +Copy these structures into project-local task and slide files as needed. Keep them compact; do not turn production work into audit reports. + +Use two lines: + +- `production-light`: normal production after the project look has been proven. Default review mode is `none`. +- `test/audit`: core-shot proving, workflow debugging, failure investigation, or explicit user-requested review. Richer records are allowed. ## Task Plan +Production-light form: + +```text +# + +## Active Scope + +- Line: production-light +- Worker profile: +- Running mode: +- Review mode: none +- Slides: +- Child session: +- Callback target: +- Write policy: append-only; overwrite allowed: no +- Status: + +## Context To Use + +## Stop Conditions +``` + +Use a longer plan only in `test/audit`. + ```text # @@ -42,90 +71,139 @@ Copy these structures into project-local task and slide files as needed. Keep th ## Task Log +Production-light form: + ```text # Task Log ## Status -## Timeline +## Dispatch / Callback Receipts -## Child Session Results +- YYYY-MM-DD HH:MM dispatch: , , , , +- YYYY-MM-DD HH:MM callback: , , , , , ## Slide Status Summary ## Open Decisions ``` -## Child Session Plan +## Production Run Status + +Use this for any production request that may continue across multiple child sessions or survive parent-session compression. Keep it current and compact. ```text -# Child Session NN Plan +# Production Run Status ## Status -## Parent Task +- Project: +- Lifecycle phase: +- Line: +- Running mode: +- Review mode: +- Package size: +- User confirmation: +- Current status: -## Parent Callback Contract +## Scope -## Assigned Slides +- Total target units: +- Completed units: +- Active child session: +- Next cursor: -## Shared Context To Read +## Callback Target -## Visual-System Files To Read +- Parent thread: +- Callback mode: -## Artifact Version / Write Policy +## Latest Callback -## Per-Slide Work Order - -## Agent Packet Paths +- Status: +- Topology: +- Resolved versions: +- Handoff path: +- Image paths: +- Problems: +- Needs user: ## Stop Conditions ``` +Field rules: + +- `Package size` records the default or the task-specific validation override. +- `Completed units` should list only units completed in the current production run, not every historical accepted unit. +- `Next cursor` is the next unresolved unit in the authorized target set. +- `User confirmation` says whether a clean callback may continue automatically or must pause for user visual confirmation. +- `Latest Callback` is updated by the parent after callback. A child session must not update project files after sending the callback. +- Do not paste prompt text, full file inventories, or image review notes into this file. + +## Child Session Plan + +Use this only when the child needs its own local plan. Normal production-light packages can proceed from the parent dispatch plus the compact handoff. + +Production-light form: + +```text +# Child Session NN Plan + +- Line: production-light +- Parent task: +- Parent callback target: +- Assigned slides: +- Running mode: +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Shared context: +- Visual-system files: +- Topology: direct-child; no Agent/SubAgent unless explicitly requested +- Stop conditions: +- Callback final action: after the child sends a successful callback, it must not write or verify additional project files. +``` + ## Child Session Handoff +Production-light form: + ```text # Child Session NN Handoff -## Status - -## Callback Status - -## Completed Slides - -## Completion Gate - -## Review Mode / Result - -## Agent Lifecycle - -## Files Written - -## Problems / Repairs Needed - -## User Decisions Needed - -## Parent Summary - -## Next Session Input +- Status: +- Topology: direct-child | test/audit-with-agent | repair +- Callback: ready-to-send-as-final-action | blocked | not required +- Review mode: none +- Resolved versions: +- Image paths: +- Problems: +- Needs user: +- Observed elapsed/context: +- Parent summary: ``` Handoff field rules: -- `Callback Status` records whether callback is required, the parent thread id, the callback mechanism, whether the callback has been sent, and confirms that `child-session-handoff.md` existed before callback. -- `Completion Gate` states `pass`, `blocked`, `failed`, or `warning`, then lists only the concise evidence needed by the parent. -- `Review Mode / Result` states whether review was required, suppressed, completed, or skipped by task rule. -- `Agent Lifecycle` lists each slide Agent and confirms terminal / closed status after return. -- `Parent Summary` is the main-session consumption surface. It should be short enough that the parent can continue without opening prompt, return, ledger, or image files unless the status is not clean. +- `Status` states `complete`, `blocked`, `failed`, or `warning`. +- `Topology` is `direct-child` for normal production-light. Use `test/audit-with-agent` only when the task explicitly uses internal slide workers. +- `Review mode` states `none` for production-light unless the task explicitly changed it. +- `Resolved versions` and `Image paths` are the main continuation surface. +- `Observed elapsed/context` is optional and should be filled only when visible to the child or reported by the user. +- `Parent summary` should be short enough that the parent can continue without opening prompt, ledger, or image files when status is clean. +- Do not include full file inventories, narrative timelines, or copied prompt text in a production-light handoff. Callback payload rules: -- Send the callback only after `child-session-handoff.md` exists. -- The callback should include status, completion gate, resolved versions, handoff path, image paths if complete, and any problems or user decisions needed. +- Send the callback only after all allowed output files, `slides/slides.md` updates, narrow required checks, and `child-session-handoff.md` are complete. +- The callback is the child session's final action. Do not update the handoff, run closeout checks, or write any project file after a successful callback. +- The production-light callback should include status, topology, resolved versions, handoff path, image paths if complete, and any problems or user decisions needed. +- Do not add a routine final verification pass, hashes, or per-file inventory to a clean production-light callback. - If callback cannot be sent, record the failure in `Callback Status`; the parent may then use the task's recorded polling fallback or user intervention path. ## Slide Agent Packet +Use this only in `test/audit`, explicit review, repair, exception handling, or a deliberately isolated single-slide run. Production-light defaults to the child session writing the brief, prompt, image, `slides.md` update, and handoff directly. + ```text # S-NN Slide Agent Packet V1 @@ -162,38 +240,73 @@ Callback payload rules: ## Slide Agent Return +Use this only when a slide agent was actually launched. Production-light direct-child packages do not write slide-agent returns. + +Test/audit or exception form: + ```text # S-NN Slide Agent Return V1 -## Status - -## Files Written - -## Completeness Gate - -## Advisor Result - -## Generation Result - -## Host Generation Evidence - -## Review Result - -## Agent Lifecycle - -## Blockers - -## Parent Summary - -## Next Action +- Status: +- Slide/version: +- Prompt: +- Image: +- Review mode: none +- Completeness gate: +- Advisor / host-native: +- Reference roles loaded: +- Agent lifecycle: terminal +- Blockers: +- Parent summary: ``` Return field rules: -- `Completeness Gate` records file existence, append-only version, no-overwrite status, reference loading status, and review presence/absence according to packet. -- `Agent Lifecycle` states that this worker is terminal / closed after this return. -- `Parent Summary` should be short and suitable for copying into the child-session handoff. +- `Completeness gate` records only the required receipt facts: expected files exist, append-only version, no-overwrite status, reference loading status, and review absence/presence according to packet. +- `Agent lifecycle` states that this worker is terminal / closed after this return. +- `Parent summary` should be short and suitable for copying into the child-session handoff. - A return with `blocked`, `failed`, or `warning` must name the exact artifact or condition that the child/main session should inspect. +- Do not add a narrative review, full file inventory, or long source audit to the return. Production-light direct-child packages should not write this return at all. + +## Slide Design + +Use this for cover pages, PPT pages, science-video pages, title cards, explainers, and any slide that needs later human layout, typography, text placement, or post-production. Ordinary MV shot slides may skip this and start at `Slide Brief`. The default `sNN-vN-design.md` is written in Chinese. If an English companion is needed, add `sNN-vN-design-en.md` beside it. + +```text +# S-NN Slide Design V1 + +## Status + +## Design Purpose + +## Discussion Summary + +## Visual Concept + +## Image Base Requirements + +## Text / Typography Plan + +## Layout / Safe Areas + +## Color Guidance + +## Post-Production Notes + +## Constraints / Avoid + +## Confirmation Gate +``` + +Design field rules: + +- Write the default `sNN-vN-design.md` in Chinese unless the user explicitly asks otherwise. +- Use `sNN-vN-design-en.md` only as an optional English companion or translation; it must not replace the Chinese default. +- `Design Purpose` says who the design is for and why a separate design file is needed. +- `Discussion Summary` captures the user-approved design direction and tradeoffs. +- `Image Base Requirements` describes what the generated image should and should not contain; if text will be added later, explicitly say the generated image must contain no text. +- `Text / Typography Plan`, `Layout / Safe Areas`, and `Color Guidance` are for human post-production, not for image generation unless the task explicitly asks to render text into the image. +- `Confirmation Gate` states what must be confirmed before `brief`, `prompt`, or image generation starts. ## Slide Brief @@ -233,7 +346,7 @@ Brief field rules: - `Risk Class` uses values such as `environment-only`, `object/body detail`, `distant human`, `medium human`, `close/face human`, `core motif`, or `repair`. - `Reference Plan` must name the active style module and any active character references. - Visible character slides must select character reference images from `visual-system/characters/character-reference-registry.md`. -- `Prompt Handoff` tells the slide agent what must be carried into the prompt. +- `Prompt Handoff` tells the child session, or the slide agent when one is used, what must be carried into the prompt. For Video Workbench image generation, the brief should treat `gpt-image-2` as prompt/spec Advisor only. It should not require or expect Garden/API execution unless the task plan explicitly changes the project policy. @@ -273,12 +386,14 @@ Artifact version / write policy rules: - Use a new slide version for every new task execution. Do not reuse an existing `vN` just because an earlier version is incomplete. - Resolve the version by scanning existing `slides/sNN/sNN-v*-*` files and choosing the next unused integer. -- A slide-agent packet must state `Overwrite allowed: no` by default. +- The child session, and any slide-agent packet when one is used, must state `Overwrite allowed: no` by default. - Before writing any output path, check whether the target file exists. If it exists and overwrite is not explicitly authorized for that exact file, stop and return an artifact-collision blocker. - Do not backfill missing brief or prompt files for an old generated image as if they were original generation inputs. If backfill is needed for analysis, label it as reconstruction / audit, not as the original prompt. ## Slide Review +Use this only in `test/audit` or explicit review tasks. Production-light default is no review file. + ```text # S-NN Slide Review V1 diff --git a/handoff/2026-06-25-on-the-road-formal-production-handoff.md b/handoff/2026-06-25-on-the-road-formal-production-handoff.md new file mode 100644 index 0000000..1eaac54 --- /dev/null +++ b/handoff/2026-06-25-on-the-road-formal-production-handoff.md @@ -0,0 +1,251 @@ +# On The Road Formal Production Handoff + +Date: 2026-06-25 + +Project: + +```text +projects/2026-06-23-在路上 +``` + +Purpose: + +Use this as the entrypoint for the next clean main session after R7. The project has completed formal production-light shot-image generation for the current 51-shot table. + +## Current Status + +The workflow test line is closed for normal production, and the planned shot-image production pass is complete. + +`project.md` now records: + +```text +Workflow phase: production-light +Lifecycle phase: assembly-closeout +``` + +Accepted production route: + +- Main session talks with the user and defines scope. +- Main session creates a clean named child session, with UI title beginning `子会话：...`. +- Child session directly writes each slide's brief and prompt, then generates images sequentially. +- Child session updates `slides/slides.md`, completes narrow required checks, writes a compact `child-session-handoff.md`, and callbacks the parent as its final action. +- After a successful callback, the child session must not write or verify additional project files. +- No Agent/SubAgent by default. +- No review file by default. +- No slide-agent packet or return receipt by default. +- No routine final completeness audit. +- Keep write-before version scanning, append-only versions, and overwrite disallowed. + +R7 evidence: + +- `session-12`: `S-26/S-27/S-28`, Agent-backed production-light, accepted by user, 25m48s / 134K tokens. +- `session-13`: `S-29/S-30/S-31`, direct child session without Agent/SubAgent, accepted by user, 10m32s / 114K tokens. + +`session-13` is the current formal production model. + +Capacity validation after R7: + +- `session-01` under `tasks/2026-06-25-production-capacity-validation`: `S-32/S-34/S-35/S-36/S-37`, direct child session without Agent/SubAgent, accepted by user. +- User observed callback at about 152K tokens / 58% context and final usage after post-callback document/check work at about 164K tokens / 63% context. +- Mechanism warning: this validation exposed a callback-order defect because the child wrote/checked after callback. Future child prompts must state that callback is the final action. +- `session-01` under `tasks/2026-06-25-8slide-compression-validation`: `S-38/S-39/S-40/S-41/S-42/S-43/S-44/S-45`, direct child session without Agent/SubAgent, accepted into production-pass by production-stage default. +- User observed final package usage at 174K tokens / 67% context and 38m37s elapsed. +- The callback-final contract was verified: no child project writes or checks were observed after the callback/final child response. +- Ordinary production-light package size is now up to 8 slides. +- `session-01` under `tasks/2026-06-25-final-production-s48-s50`: `S-48/S-49/S-50`, direct child session without Agent/SubAgent, completed with `problems: none` and `needs_user: none`. +- Parent verification found expected brief/prompt/image files present, exact review files absent, `slides/slides.md` updated, and no post-callback child writes. + +## New Session Load Order + +For a cleanup / discussion-only new session, load in this order: + +```text +AGENTS.md +handoff/2026-06-25-on-the-road-formal-production-handoff.md +projects/2026-06-23-在路上/project.md +projects/2026-06-23-在路上/execution-plan.md +projects/2026-06-23-在路上/slides/slides.md +``` + +For a session that may dispatch production image generation, then load: + +```text +VIDEO_WORKBENCH.md +docs/workflows/project-lifecycle.md +docs/workflows/slide-task-orchestration.md +docs/workflows/slide-task-templates.md +projects/2026-06-23-在路上/visual-system/visual-system.md +projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md +projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md +``` + +For the actual assigned slide scope, load only the relevant source rows and files: + +```text +projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md +projects/2026-06-23-在路上/intake/阶段五-核心单镜.md # only for S-10/S-46/S-51 or if referenced +projects/2026-06-23-在路上/slides/sNN/ +``` + +Do not read old session directories, old handoffs, or the external discussion record unless investigating a specific inconsistency. This handoff supersedes the R7 governance-simplification handoff for normal production. + +## If The User Directly Says "生图" + +First classify the request: + +1. Clear slide scope: proceed with that named scope. +2. Clear natural-language scope: map it to slide IDs using `slides/slides.md`, then confirm only if mapping is ambiguous. +3. No scope, just "继续生图" or equivalent: use the "Default Next Scope" section below. +4. Repair / local problem / quality concern: ask whether to use `production-light` retry or `test/audit` override. + +Do not start generation without an explicit slide scope unless the user clearly authorizes "continue from the current next production scope". + +## Default Next Scope + +Accepted / production-pass now includes the current 51-shot table. Key late-production coverage: + +```text +S-26/S-27/S-28/S-29/S-30/S-31/S-32/S-33/S-34/S-35/S-36/S-37/S-38/S-39/S-40/S-41/S-42/S-43/S-44/S-45/S-46/S-47/S-48/S-49/S-50/S-51 +``` + +The remaining not-started story-order items are: + +```text +none +``` + +Reason: + +- `S-32/s32-v2`, `S-34/s34-v2`, `S-35/s35-v1`, `S-36/s36-v1`, and `S-37/s37-v1` were accepted by user in the capacity validation package. +- `S-33` is already accepted as `s33-v2`; skip by default. +- `S-38` through `S-45` are accepted into the production pass after the 8-slide compression validation. +- `S-46`, `S-47`, and `S-51` are already accepted anchor/trial outputs. +- `S-48` through `S-50` were completed in the final production direct-child package. + +If the user says "直接继续正式生产" with no other scope, do not dispatch a new image-production package. Report that the current shot-image table is complete and ask whether to proceed to assembly / export / closeout or to name a slide for append-only repair. + +```text +no remaining default production scope +``` + +For append-only repair/retry, the child session must still scan each slide folder before writing; any expected versions are not authority. + +## Scope Rules + +Use `slides/slides.md` as the source of truth for production scope. + +Closed by default: + +- status `accepted`; +- status `production-pass`; +- rows explicitly recorded as user accepted in current project/task records. + +Open by default: + +- status `not started`; +- status `generated` without user acceptance; +- broad-run candidate only; +- incomplete / paused evidence; +- rejected historical evidence; +- repair requested by user. + +Important cleanup note: + +Some older rows still use wording such as `generated / session-07 complete` even where project/task records say the user accepted the images. The next cleanup pass should normalize these statuses before large-scale batching. Until then, for any ambiguous row, prefer the latest explicit user-acceptance record in `project.md`, `execution-plan.md`, `task-log.md`, and `slides.md`. + +Default child-session package size: + +```text +maximum 8 ordinary production slides +``` + +Use 1-3 slides for identity-heavy, repair-heavy, core-risk, or context-expensive scopes. Split larger scopes into multiple child sessions unless the user explicitly authorizes a larger package. + +## Production Dispatch Defaults + +When dispatching a production-light package, create one clean child session. + +Title format: + +```text +子会话：在路上 S48-S50 production-light +``` + +Use a concise parent instruction with these fields: + +```text +Project: projects/2026-06-23-在路上 +Line: production-light +Topology: direct child session; no Agent/SubAgent +Worker profile: video-slide-production +Running mode: prompt and generate +Review mode: none +Scope: +Callback target: +Write policy: append-only; overwrite allowed: no +Required source of truth: project.md, execution-plan.md, slides/slides.md, visual-system/ +Output: brief, prompt, image, slides.md update, compact child-session-handoff.md +Callback payload: status, topology, resolved versions, handoff path, image paths, problems, needs_user +Callback order: callback is the child session's final action after all allowed writes, narrow checks, and handoff work; do not write or verify project files after a successful callback +``` + +The child must: + +- scan existing `slides/sNN/sNN-v*-*` files before choosing each version; +- stop on artifact collision instead of overwriting; +- create no `sNN-vN-review.md` unless the parent explicitly changes review mode; +- create no packet/return files unless the parent explicitly switches to test/audit, repair, or Agent-backed mode; +- use `gpt-image-2` Advisor / host-native workflow; +- load real visible-context image references before generation when identity or reference control matters; +- keep prompts compact and generation-oriented; +- update `slides/slides.md` for completed outputs; +- write compact handoff before callback; +- send callback as the final child action and perform no project writes/checks afterward. + +The parent must: + +- not poll the child thread routinely; +- not audit prompt/image/ledger details after a clean callback; +- inspect details only on `blocked`, `failed`, `warning`, collision, explicit user audit, or ambiguous scope/status. + +## Test / Audit Override + +Use `test/audit` only when: + +- the user asks to investigate a workflow problem; +- a generated image fails and needs diagnosis rather than a simple append-only retry; +- a new core visual rule needs proving; +- external review / local review / Agent isolation is explicitly requested; +- duplicate worker, collision, missing reference, or tool failure occurs. + +In `test/audit`, Agent/SubAgent, packet/return files, local review, and fuller completion gates are allowed. They are not the formal production default. + +## Current Known Boundaries + +- Do not invoke CCPE. +- Do not ask GPT V2 for final prompts. +- Do not call MiniMax MCP or external vision/review services unless explicitly authorized for the task. +- Do not regenerate accepted slides by default. +- Do not reuse broad-run candidate images as accepted finals unless the user explicitly accepts them. +- Do not backfill old broad-run prompt/brief files as if they were original generation inputs. +- Do not overwrite existing artifacts. + +## Cleanup Topics For The Next Session + +Likely useful cleanup / discussion items before or during production: + +- Normalize old `slides.md` statuses where user acceptance is recorded elsewhere but row status still says `generated`. +- Decide whether broad-run `s32-v1` and `s34-v1` should be accepted as candidates or superseded by formal append-only production versions. +- Consider whether `S-15/S-17/S-18` are merely generated historical evidence or should be reviewed/accepted/retried before final assembly. +- Keep R7 production-light rules; do not reopen the Agent/default-governance debate unless a new failure appears. + +## Stop Conditions + +Stop and ask the user if: + +- the requested slide scope cannot be mapped cleanly; +- the next scope would include an already accepted slide; +- a target output path already exists; +- required character/style references cannot be loaded; +- the user request implies review/audit but does not say which line to use; +- the child callback reports `blocked`, `failed`, or `warning`. diff --git a/projects/2026-06-23-在路上/execution-plan.md b/projects/2026-06-23-在路上/execution-plan.md index 397d2b8..bb8b00e 100644 --- a/projects/2026-06-23-在路上/execution-plan.md +++ b/projects/2026-06-23-在路上/execution-plan.md @@ -9,14 +9,14 @@ ## Current Execution Focus -Prepare the execution base, not the full MV. +Shot-image production is complete for the current 51-shot table. Next work is assembly / export / closeout when requested. Immediate focus: -1. Build and validate the character anchor board. -2. Build and validate the style anchor board. -3. Use both anchors in a small visual test package. -4. Only after tests pass, expand into broader shot production. +1. Do not dispatch more production image packages unless the user requests append-only repair/retry for named slides. +2. Use direct child sessions for normal production-light packages. +3. Keep review `none` unless the user explicitly asks for review. +4. Use test/audit routing only for repairs, workflow debugging, or explicit investigation. ## Materialization Targets @@ -71,12 +71,17 @@ Round 1 success means these are true: - R5 brief-ownership test completed: clean child session `019ef907-16b7-7a43-9924-227892aa62d3` generated accepted `s07-v2`, `s08-v3`, and `s09-v2` with child-owned brief/packet before worker launch for successful outputs. `s08-v2` remains incomplete transport-failure evidence and was retried append-only as `s08-v3`. - Package-size trials completed through a user-authorized 4-slide package. `session-07` produced accepted `s11-v2`, `s12-v2`, and `s13-v2` with observed context 157K and elapsed time 30m5s. `session-08` mechanically completed `s15-v2` through `s18-v2`, but `s16-v2` failed visual acceptance because it did not achieve `横向大远景，路从画面一侧切向远方，人物极小` and was too similar to `s15`. - R6 Phase 1 notification/wakeup test completed and user accepted the image: clean child session `019ef9bc-fe96-75b3-b902-5d5ef95feca3` generated append-only `s16-v3` and sent a callback to parent thread `019ef9b5-cc39-7a42-91b1-f9b0852dc95e`; parent post-processing used local handoff/file evidence without polling the child thread. +- R7 production-light simplification test completed: `session-12` generated `s26-v2`, `s27-v2`, and `s28-v2` with review `none`; user accepted all three images and observed 25m48s / 134K tokens. +- R7 direct-child production-light test completed: `session-13` generated `s29-v2`, `s30-v2`, and `s31-v2` directly without Agent/SubAgent; user accepted all three images and observed 10m32s / 114K tokens. This is now the formal production default. +- Production capacity validation completed: direct child session `019efc24-96a4-7582-bb05-914b82a726d5` generated `s32-v2`, `s34-v2`, `s35-v1`, `s36-v1`, and `s37-v1` without Agent/SubAgent or review files; user accepted all five images. User observed callback at about 152K tokens / 58% context and final usage after post-callback document/check work at about 164K tokens / 63% context. +- 8-slide compression validation completed: direct child session `019efc48-d4af-7972-8646-6d04778dfb2c` generated `s38-v1` through `s45-v1` without Agent/SubAgent or review files; user promoted the package to production-pass by production-stage default. User observed 174K tokens / 67% context / 38m37s elapsed. Ordinary production-light packages may now use up to 8 slides, but child callback must be the final action. +- Final production package completed: direct child session `019efc78-e0f1-7601-aaec-bab4f3e6f99d` generated `s48-v1`, `s49-v1`, and `s50-v1` without Agent/SubAgent or review files; child reported `complete`, `problems: none`, `needs_user: none`; parent verification found expected files present, exact review files absent, and no post-callback child writes. ## Active Batch -R6 Phase 1 callback-only `S-16` repair is accepted. R6 Phase 2 first 4-shot attempt was paused by the user after the `S-20` worker attempted unauthorized MiniMax/MCP image understanding and hung; `S-19/s19-v2` is accepted. R6 Phase 2 retry completed in clean `session-11` for `S-20/S-22/S-23/S-24`; user accepted `s20-v3`, `s22-v3`, `s23-v3`, and `s24-v2`. +No active child production package is running. R7 production-light outputs `s26-v2`, `s27-v2`, `s28-v2`, `s29-v2`, `s30-v2`, and `s31-v2`, capacity-validation outputs `s32-v2`, `s34-v2`, `s35-v1`, `s36-v1`, and `s37-v1`, 8-slide validation outputs `s38-v1` through `s45-v1`, and final production outputs `s48-v1` through `s50-v1` are accepted into production-pass. Direct-child, no-Agent topology remains the current production route for future repair/retry packages. -No single-shot gate is active. The five-shot visual trial is accepted. The character-consistency workflow test is visually accepted for S-21, S-14, S-25, S-33, S-06, S-07, S-08, S-09, S-11, S-12, S-13, and S-16/s16-v3. R4 images for S-02, S-03, and S-04 are visually accepted but remain a mechanism-warning run because child-session brief ownership did not match the desired boundary. R5 successful outputs passed the child-owned brief/packet rule. `S-16/s16-v2` remains rejected historical evidence. +No single-shot gate is active. The five-shot visual trial is accepted. The character-consistency workflow test is visually accepted for S-21, S-14, S-25, S-33, S-06, S-07, S-08, S-09, S-11, S-12, S-13, S-16/s16-v3, S-20/s20-v3, S-22/s22-v3, S-23/s23-v3, S-24/s24-v2, S-26 through S-32, and S-34 through S-50. R4 images for S-02, S-03, and S-04 are visually accepted but remain a mechanism-warning run because child-session brief ownership did not match the desired boundary. `S-16/s16-v2` remains rejected historical evidence. Current broad-run issue report: @@ -92,15 +97,15 @@ slides/broad-generation-self-audit-v1.md | Character board size | 16:9 page, shallow gray-white background. | | Character board scope | MVP can start with front/side/back full body plus half-body state; full version should add walking side/back and prop breakdowns. | | Style board scope | MVP one-page A/B/C/D board; C should be marked as repaired / accepted. | -| Full shot production | Wait for the next user production decision after R6 Phase 2 retry acceptance. Current default route: per-shot brief/prompt trace restored, visible character shots use actual references, child sessions may handle up to 3 slides by default unless explicitly overridden, clean user-visible child sessions for orchestration, visible child writes brief/packet before worker launch, internal one-shot SubAgents for slide work, one slide/version has exactly one owning worker, child sessions perform the completion gate before handoff, callback-only parent wakeup is proven at single-slide and 4-shot package production scope, and MiniMax MCP/external vision remains off unless explicitly authorized. | +| Full shot production | Authorized for production-light. Current default route: named clean user-visible child sessions (`子会话：...`) handle up to 8 ordinary production slides directly, write brief/prompt/image/slide status/handoff, use append-only version scanning with overwrite disallowed, suppress review files, skip Agent/SubAgent, skip packet/return artifacts, skip routine final completeness audit, and callback the parent as the final child action when complete or blocked. Use 1-3 slides for identity-heavy, repair-heavy, core-risk, or context-expensive scopes. MiniMax MCP/external vision remains off unless explicitly authorized. | ## Next Decision -R6 next decision: +Next decision: -- User has accepted `S-20/s20-v3`, `S-22/s22-v3`, `S-23/s23-v3`, and `S-24/s24-v2`; wait for the next production scope decision. -- Keep callback-only behavior: parent dispatches and exits; child writes compact handoff and callbacks parent; parent reads local handoff by default and inspects details only on warning/blocked/failed. +- Shot-image production is complete for the current 51-shot table; there is no remaining not-started production scope. +- Keep callback-only behavior: parent dispatches and exits; child completes all writes/checks/handoff work first, then callbacks parent as the final action. Parent reads local handoff by default and inspects details only on warning/blocked/failed. - Keep the quality route: `gpt-image-2` Advisor prompt/spec, active style master, active character registry when needed, and visible reference images loaded before generation. -- Keep the brief ownership rule: the visible child session writes and confirms `sNN-vN-brief.md` and the slide-agent packet before launching a one-shot slide worker. +- Keep the production-light topology: the visible child session writes `sNN-vN-brief.md`, `sNN-vN-prompt.md`, `sNN-vN-image.png`, `slides.md` updates, and a compact handoff directly. Agent/SubAgent and packet/return files are opt-in for test/audit, repair, exception, or explicit isolated-slide work. - Keep append-only retry behavior: incomplete outputs such as `s08-v2` remain historical evidence and retries use the next unused version. -- Keep the duplicate-worker guard: one successful slide/version must have exactly one owning worker and one prompt/image/return target set. +- Keep the write-before version scan: one successful slide/version should have one brief/prompt/image target set, and existing files must not be overwritten without explicit authorization. diff --git a/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/fix-title-batch-report.md b/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/fix-title-batch-report.md new file mode 100644 index 0000000..aac8e58 --- /dev/null +++ b/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/fix-title-batch-report.md @@ -0,0 +1,5 @@ +# Fix Title Batch Report + +| Source | Mode | Heading Map | Plan | Fixed | Report | +| --- | --- | --- | --- | --- | --- | +| C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\intake\阶段五-封面设计骨架.md | artifact | C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\intake-repair\fix-title-2026-06-25-cover-design\阶段五-封面设计骨架.heading-map.json | C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\intake-repair\fix-title-2026-06-25-cover-design\阶段五-封面设计骨架.heading-plan.json | C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\intake-repair\fix-title-2026-06-25-cover-design\阶段五-封面设计骨架.fixed.md | C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\intake-repair\fix-title-2026-06-25-cover-design\阶段五-封面设计骨架.heading-report.md | diff --git a/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.fixed.md b/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.fixed.md new file mode 100644 index 0000000..ca5233b --- /dev/null +++ b/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.fixed.md @@ -0,0 +1,340 @@ +# 《在路上》MV 封面设计骨架 V1 + +## 0. 封面定位 + +这张封面不是人物海报，也不是普通风景图。 + +它的任务是： + +> **让人一眼知道：这是一首关于继续走的歌。** +> 路很长，人很小，泥很深，但泥里有星，远方有火。 + +封面要承担整支 MV 的视觉总钥匙： + +* 看见路； +* 看见行者； +* 看见泥泞； +* 看见星空意象； +* 看见远方火光； +* 但不能变成人物写真。 + +--- + +## 1. 推荐主方向 + +### 主封面方向：**泥路尽头，泥水有星，远方有火** + +这是我最推荐的封面方向。 + +因为它能同时压缩全片 4 个核心母题： + +| 母题 | 封面中如何出现 | +| -- | ------------- | +| 路 | 一条泥泞长路延伸向远方 | +| 人 | 一个很小的中年行者背影 | +| 星 | 前景泥水中有破碎星空倒影 | +| 火 | 远方黑暗里有一簇或两簇小火 | + +这比单纯大远景更有记忆点， +也比单纯“泥泞里的星空”更适合作为封面，因为它多了人物和方向感。 + +--- + +## 2. 封面核心画面 + +### 画面描述 + +16:9 横版封面。 + +夜色或黎明前的泥泞道路。 +前景是一片泥水洼，水面里隐约倒映破碎星空。 +中远景里，一个 50 岁左右的行者背着旧包，穿深色长外套，沿着泥路向远方走。 +人物很小，不露脸。 +远处黑暗里有一簇微弱火光，可能还有第二簇更远的火点。 + +整体感觉： + +> 不是他已经抵达。 +> 是他还在路上。 +> 但这一次，他知道泥里有星，远方有火。 + +--- + +## 3. 构图骨架 + +### 画面三层结构 + +#### 第一层：前景 + +**泥泞水洼 + 破碎星空倒影** + +功能： +承载全片灵魂意象。 + +注意： +水洼不能太脏到没有美感，也不能太梦幻。 +要有真实泥感，也要有星光。 + +--- + +#### 第二层：中景 + +**小人物背影 + 旧背包 + 长外套** + +功能： +建立主角身份，但不让人物抢戏。 + +人物占比建议： + +> 画面高度的 1/6 左右，甚至更小。 + +重点不是脸，而是： + +* 背影； +* 外套轮廓； +* 背包； +* 行走姿态。 + +--- + +#### 第三层：远景 + +**长路 + 远方火光 + 暗色天空** + +功能： +给封面方向感和结尾余韵。 + +远处火光要小。 +不能像篝火晚会，也不能像露营宣传片。 +它只是告诉观众： + +> 远方有人。 + +--- + +## 4. 封面版式建议 + +### 主标题 + +标题： + +> 在路上 + +建议放置位置： + +#### 推荐位置 A：画面左上或右上 + +优点： +不挡住前景泥水星空，也不挡住人物。 + +#### 推荐位置 B：画面下方偏左 + +优点： +更像音乐封面，但要避免压住水洼星空。 + +我更推荐： + +> **标题放在画面左上 / 右上，保持留白和克制。** + +--- + +### 字体气质 + +标题字体不要太现代科技感，也不要太文艺花体。 + +建议气质： + +* 有手写感但不潦草； +* 有粗粝感但不破败； +* 像民谣摇滚封面； +* 温暖、克制、带一点风尘。 + +可以考虑： + +* 粗一点的手写体； +* 有笔触感的宋黑混合风； +* 简洁但略有旧感的中文标题字。 + +--- + +### 副标题 / 信息 + +封面不建议放太多字。 + +最多可以有一行小字： + +> Acoustic Indie Road Rock + +或： + +> 泥泞里，有星空 + +但我更建议主封面只保留： + +> **在路上** + +因为这张图的意象已经够强。 +副标题容易削弱高级感。 + +--- + +## 5. 色彩系统 + +### 主色调 + +* 深蓝黑； +* 冷灰； +* 泥土褐； +* 暗卡其； +* 湿地反光的银蓝灰。 + +### 情绪色 + +* 远方火光的暖橙； +* 水洼星点的冷白； +* 可能有一点黎明前的暗蓝。 + +### 色彩原则 + +> 大面积冷暗，小面积暖光。 +> 泥是现实，星是精神，火是同行。 + +不要做成： + +* 金色励志大片； +* 旅游风景片； +* 商业橙蓝大片； +* 赛博蓝紫科技感； +* 过度唯美的日落封面。 + +--- + +## 6. 人物处理规则 + +封面人物必须遵守： + +* 不露脸； +* 不居中大特写； +* 不做英雄姿态； +* 不张开双臂； +* 不回头看镜头； +* 不像旅行博主； +* 不像摇滚明星海报。 + +人物应是： + +> 环境中的一个行者锚点。 + +最好是： + +* 背影； +* 稍偏画面一侧； +* 正在走； +* 外套下摆和背包轮廓清楚； +* 身体有疲惫感，但不塌。 + +--- + +## 7. 封面信息层级 + +封面第一眼看到： + +> 夜路 / 泥水 / 星光 / 远方火。 + +第二眼看到： + +> 有一个人还在走。 + +第三眼看到： + +> 歌名《在路上》。 + +不要让第一眼变成： + +> 一个中年男人的背影照。 + +--- + +## 8. 可选封面方案 + +### 方案 A：主推版 + +#### **泥水星空 + 小人物 + 远方火光** + +适合做正式 MV 封面。 + +优势： +最完整地压缩全片精神。 + +风险： +泥水星空必须处理好，否则会脏或不够诗性。 + +--- + +### 方案 B：风景长路版 + +#### **清晨泥泞长路 + 小人物背影** + +适合作为更克制的备用封面。 + +优势： +构图稳，风景叙事感强。 + +风险： +记忆点可能不如 A，容易只是“公路片封面”。 + +--- + +### 方案 C：夜路篝火版 + +#### **夜色长路 + 小火 + 远方火点** + +适合作为情绪更温暖的备用封面。 + +优势： +结尾感强，同行者意象明显。 + +风险： +容易偏露营，必须压住火的热闹感。 + +--- + +## 9. 我建议的最终封面方向 + +优先做： + +> **方案 A：泥水星空 + 小人物 + 远方火光** + +因为它最像《在路上》的精神浓缩版。 + +一句话定义： + +> **一个人走在泥泞夜路上，脚下是破碎星空，远方有微弱火光。** + +这张封面如果成立，观众不用看歌词，也能感到： + +> 这不是胜利的歌。 +> 这是继续走的歌。 + +--- + +## 10. 封面验收标准 + +封面完成后，强哥可以按这 8 条审： + +1. 第一眼是不是看到路和世界，而不是脸？ +2. 人物是否足够小？ +3. 泥水星空是否明确？ +4. 远方火光是否克制？ +5. 有没有避免露营感？ +6. 有没有避免人物写真感？ +7. 歌名字体是否克制、有风尘感？ +8. 缩小成视频缩略图后，是否仍能看出“路、火、人、星”？ + +--- + +## 11. 封面一句话标准 + +最终封面应该守住这一句： + +> **泥泞是真的，星空也是真的；人很小，但还在走，远方还有火。** diff --git a/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.heading-map.json b/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.heading-map.json new file mode 100644 index 0000000..717110c --- /dev/null +++ b/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.heading-map.json @@ -0,0 +1,171 @@ +{ + "source": "C:\\Users\\wangq\\Documents\\Codex\\video-workbench\\projects\\2026-06-23-在路上\\intake\\阶段五-封面设计骨架.md", + "heading_count": 33, + "headings": [ + { + "line": 1, + "level": 1, + "text": "《在路上》MV 封面设计骨架 V1" + }, + { + "line": 3, + "level": 2, + "text": "0. 封面定位" + }, + { + "line": 23, + "level": 1, + "text": "1. 推荐主方向" + }, + { + "line": 25, + "level": 2, + "text": "主封面方向：**泥路尽头，泥水有星，远方有火**" + }, + { + "line": 43, + "level": 1, + "text": "2. 封面核心画面" + }, + { + "line": 45, + "level": 2, + "text": "画面描述" + }, + { + "line": 63, + "level": 1, + "text": "3. 构图骨架" + }, + { + "line": 65, + "level": 2, + "text": "画面三层结构" + }, + { + "line": 67, + "level": 3, + "text": "第一层：前景" + }, + { + "line": 80, + "level": 3, + "text": "第二层：中景" + }, + { + "line": 100, + "level": 3, + "text": "第三层：远景" + }, + { + "line": 115, + "level": 1, + "text": "4. 封面版式建议" + }, + { + "line": 117, + "level": 2, + "text": "主标题" + }, + { + "line": 125, + "level": 3, + "text": "推荐位置 A：画面左上或右上" + }, + { + "line": 130, + "level": 3, + "text": "推荐位置 B：画面下方偏左" + }, + { + "line": 141, + "level": 2, + "text": "字体气质" + }, + { + "line": 160, + "level": 2, + "text": "副标题 / 信息" + }, + { + "line": 181, + "level": 1, + "text": "5. 色彩系统" + }, + { + "line": 183, + "level": 2, + "text": "主色调" + }, + { + "line": 191, + "level": 2, + "text": "情绪色" + }, + { + "line": 197, + "level": 2, + "text": "色彩原则" + }, + { + "line": 212, + "level": 1, + "text": "6. 人物处理规则" + }, + { + "line": 238, + "level": 1, + "text": "7. 封面信息层级" + }, + { + "line": 258, + "level": 1, + "text": "8. 可选封面方案" + }, + { + "line": 260, + "level": 2, + "text": "方案 A：主推版" + }, + { + "line": 262, + "level": 3, + "text": "**泥水星空 + 小人物 + 远方火光**" + }, + { + "line": 274, + "level": 2, + "text": "方案 B：风景长路版" + }, + { + "line": 276, + "level": 3, + "text": "**清晨泥泞长路 + 小人物背影**" + }, + { + "line": 288, + "level": 2, + "text": "方案 C：夜路篝火版" + }, + { + "line": 290, + "level": 3, + "text": "**夜色长路 + 小火 + 远方火点**" + }, + { + "line": 302, + "level": 1, + "text": "9. 我建议的最终封面方向" + }, + { + "line": 321, + "level": 1, + "text": "10. 封面验收标准" + }, + { + "line": 336, + "level": 1, + "text": "11. 封面一句话标准" + } + ] +} diff --git a/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.heading-plan.json b/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.heading-plan.json new file mode 100644 index 0000000..b9e57ba --- /dev/null +++ b/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.heading-plan.json @@ -0,0 +1,194 @@ +{ + "mode": "artifact", + "source": "C:\\Users\\wangq\\Documents\\Codex\\video-workbench\\projects\\2026-06-23-在路上\\intake\\阶段五-封面设计骨架.md", + "output": "C:\\Users\\wangq\\Documents\\Codex\\video-workbench\\projects\\2026-06-23-在路上\\intake-repair\\fix-title-2026-06-25-cover-design\\阶段五-封面设计骨架.fixed.md", + "report": "C:\\Users\\wangq\\Documents\\Codex\\video-workbench\\projects\\2026-06-23-在路上\\intake-repair\\fix-title-2026-06-25-cover-design\\阶段五-封面设计骨架.heading-report.md", + "edits": [ + { + "line": 23, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + }, + { + "line": 25, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 23." + }, + { + "line": 43, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + }, + { + "line": 45, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 43." + }, + { + "line": 63, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + }, + { + "line": 65, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 63." + }, + { + "line": 67, + "from_level": 3, + "to_level": 4, + "reason": "Subsection under the repaired child heading at line 65." + }, + { + "line": 80, + "from_level": 3, + "to_level": 4, + "reason": "Subsection under the repaired child heading at line 65." + }, + { + "line": 100, + "from_level": 3, + "to_level": 4, + "reason": "Subsection under the repaired child heading at line 65." + }, + { + "line": 115, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + }, + { + "line": 117, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 115." + }, + { + "line": 125, + "from_level": 3, + "to_level": 4, + "reason": "Subsection under the repaired child heading at line 117." + }, + { + "line": 130, + "from_level": 3, + "to_level": 4, + "reason": "Subsection under the repaired child heading at line 117." + }, + { + "line": 141, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 115." + }, + { + "line": 160, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 115." + }, + { + "line": 181, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + }, + { + "line": 183, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 181." + }, + { + "line": 191, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 181." + }, + { + "line": 197, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 181." + }, + { + "line": 212, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + }, + { + "line": 238, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + }, + { + "line": 258, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + }, + { + "line": 260, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 258." + }, + { + "line": 262, + "from_level": 3, + "to_level": 4, + "reason": "Subsection under the repaired child heading at line 260." + }, + { + "line": 274, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 258." + }, + { + "line": 276, + "from_level": 3, + "to_level": 4, + "reason": "Subsection under the repaired child heading at line 274." + }, + { + "line": 288, + "from_level": 2, + "to_level": 3, + "reason": "Child heading under the repaired section at line 258." + }, + { + "line": 290, + "from_level": 3, + "to_level": 4, + "reason": "Subsection under the repaired child heading at line 288." + }, + { + "line": 302, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + }, + { + "line": 321, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + }, + { + "line": 336, + "from_level": 1, + "to_level": 2, + "reason": "Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it." + } + ] +} diff --git a/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.heading-report.md b/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.heading-report.md new file mode 100644 index 0000000..0481945 --- /dev/null +++ b/projects/2026-06-23-在路上/intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.heading-report.md @@ -0,0 +1,41 @@ +# Semantic Heading Repair Report + +- Source: `C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\intake\阶段五-封面设计骨架.md` +- Output: `C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\intake-repair\fix-title-2026-06-25-cover-design\阶段五-封面设计骨架.fixed.md` +- Plan: `C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\intake-repair\fix-title-2026-06-25-cover-design\阶段五-封面设计骨架.heading-plan.json` +- Mode: `artifact` +- Changed headings: 31 + +| Line | From | To | Heading | Reason | +| --- | --- | --- | --- | --- | +| 23 | 1 | 2 | 1. 推荐主方向 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | +| 25 | 2 | 3 | 主封面方向：**泥路尽头，泥水有星，远方有火** | Child heading under the repaired section at line 23. | +| 43 | 1 | 2 | 2. 封面核心画面 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | +| 45 | 2 | 3 | 画面描述 | Child heading under the repaired section at line 43. | +| 63 | 1 | 2 | 3. 构图骨架 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | +| 65 | 2 | 3 | 画面三层结构 | Child heading under the repaired section at line 63. | +| 67 | 3 | 4 | 第一层：前景 | Subsection under the repaired child heading at line 65. | +| 80 | 3 | 4 | 第二层：中景 | Subsection under the repaired child heading at line 65. | +| 100 | 3 | 4 | 第三层：远景 | Subsection under the repaired child heading at line 65. | +| 115 | 1 | 2 | 4. 封面版式建议 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | +| 117 | 2 | 3 | 主标题 | Child heading under the repaired section at line 115. | +| 125 | 3 | 4 | 推荐位置 A：画面左上或右上 | Subsection under the repaired child heading at line 117. | +| 130 | 3 | 4 | 推荐位置 B：画面下方偏左 | Subsection under the repaired child heading at line 117. | +| 141 | 2 | 3 | 字体气质 | Child heading under the repaired section at line 115. | +| 160 | 2 | 3 | 副标题 / 信息 | Child heading under the repaired section at line 115. | +| 181 | 1 | 2 | 5. 色彩系统 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | +| 183 | 2 | 3 | 主色调 | Child heading under the repaired section at line 181. | +| 191 | 2 | 3 | 情绪色 | Child heading under the repaired section at line 181. | +| 197 | 2 | 3 | 色彩原则 | Child heading under the repaired section at line 181. | +| 212 | 1 | 2 | 6. 人物处理规则 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | +| 238 | 1 | 2 | 7. 封面信息层级 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | +| 258 | 1 | 2 | 8. 可选封面方案 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | +| 260 | 2 | 3 | 方案 A：主推版 | Child heading under the repaired section at line 258. | +| 262 | 3 | 4 | **泥水星空 + 小人物 + 远方火光** | Subsection under the repaired child heading at line 260. | +| 274 | 2 | 3 | 方案 B：风景长路版 | Child heading under the repaired section at line 258. | +| 276 | 3 | 4 | **清晨泥泞长路 + 小人物背影** | Subsection under the repaired child heading at line 274. | +| 288 | 2 | 3 | 方案 C：夜路篝火版 | Child heading under the repaired section at line 258. | +| 290 | 3 | 4 | **夜色长路 + 小火 + 远方火点** | Subsection under the repaired child heading at line 288. | +| 302 | 1 | 2 | 9. 我建议的最终封面方向 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | +| 321 | 1 | 2 | 10. 封面验收标准 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | +| 336 | 1 | 2 | 11. 封面一句话标准 | Artifact mode: keep line 1 as the sole document title; this is a peer top-level section under it. | diff --git a/projects/2026-06-23-在路上/intake/阶段五-封面设计骨架.md b/projects/2026-06-23-在路上/intake/阶段五-封面设计骨架.md new file mode 100644 index 0000000..ca5233b --- /dev/null +++ b/projects/2026-06-23-在路上/intake/阶段五-封面设计骨架.md @@ -0,0 +1,340 @@ +# 《在路上》MV 封面设计骨架 V1 + +## 0. 封面定位 + +这张封面不是人物海报，也不是普通风景图。 + +它的任务是： + +> **让人一眼知道：这是一首关于继续走的歌。** +> 路很长，人很小，泥很深，但泥里有星，远方有火。 + +封面要承担整支 MV 的视觉总钥匙： + +* 看见路； +* 看见行者； +* 看见泥泞； +* 看见星空意象； +* 看见远方火光； +* 但不能变成人物写真。 + +--- + +## 1. 推荐主方向 + +### 主封面方向：**泥路尽头，泥水有星，远方有火** + +这是我最推荐的封面方向。 + +因为它能同时压缩全片 4 个核心母题： + +| 母题 | 封面中如何出现 | +| -- | ------------- | +| 路 | 一条泥泞长路延伸向远方 | +| 人 | 一个很小的中年行者背影 | +| 星 | 前景泥水中有破碎星空倒影 | +| 火 | 远方黑暗里有一簇或两簇小火 | + +这比单纯大远景更有记忆点， +也比单纯“泥泞里的星空”更适合作为封面，因为它多了人物和方向感。 + +--- + +## 2. 封面核心画面 + +### 画面描述 + +16:9 横版封面。 + +夜色或黎明前的泥泞道路。 +前景是一片泥水洼，水面里隐约倒映破碎星空。 +中远景里，一个 50 岁左右的行者背着旧包，穿深色长外套，沿着泥路向远方走。 +人物很小，不露脸。 +远处黑暗里有一簇微弱火光，可能还有第二簇更远的火点。 + +整体感觉： + +> 不是他已经抵达。 +> 是他还在路上。 +> 但这一次，他知道泥里有星，远方有火。 + +--- + +## 3. 构图骨架 + +### 画面三层结构 + +#### 第一层：前景 + +**泥泞水洼 + 破碎星空倒影** + +功能： +承载全片灵魂意象。 + +注意： +水洼不能太脏到没有美感，也不能太梦幻。 +要有真实泥感，也要有星光。 + +--- + +#### 第二层：中景 + +**小人物背影 + 旧背包 + 长外套** + +功能： +建立主角身份，但不让人物抢戏。 + +人物占比建议： + +> 画面高度的 1/6 左右，甚至更小。 + +重点不是脸，而是： + +* 背影； +* 外套轮廓； +* 背包； +* 行走姿态。 + +--- + +#### 第三层：远景 + +**长路 + 远方火光 + 暗色天空** + +功能： +给封面方向感和结尾余韵。 + +远处火光要小。 +不能像篝火晚会，也不能像露营宣传片。 +它只是告诉观众： + +> 远方有人。 + +--- + +## 4. 封面版式建议 + +### 主标题 + +标题： + +> 在路上 + +建议放置位置： + +#### 推荐位置 A：画面左上或右上 + +优点： +不挡住前景泥水星空，也不挡住人物。 + +#### 推荐位置 B：画面下方偏左 + +优点： +更像音乐封面，但要避免压住水洼星空。 + +我更推荐： + +> **标题放在画面左上 / 右上，保持留白和克制。** + +--- + +### 字体气质 + +标题字体不要太现代科技感，也不要太文艺花体。 + +建议气质： + +* 有手写感但不潦草； +* 有粗粝感但不破败； +* 像民谣摇滚封面； +* 温暖、克制、带一点风尘。 + +可以考虑： + +* 粗一点的手写体； +* 有笔触感的宋黑混合风； +* 简洁但略有旧感的中文标题字。 + +--- + +### 副标题 / 信息 + +封面不建议放太多字。 + +最多可以有一行小字： + +> Acoustic Indie Road Rock + +或： + +> 泥泞里，有星空 + +但我更建议主封面只保留： + +> **在路上** + +因为这张图的意象已经够强。 +副标题容易削弱高级感。 + +--- + +## 5. 色彩系统 + +### 主色调 + +* 深蓝黑； +* 冷灰； +* 泥土褐； +* 暗卡其； +* 湿地反光的银蓝灰。 + +### 情绪色 + +* 远方火光的暖橙； +* 水洼星点的冷白； +* 可能有一点黎明前的暗蓝。 + +### 色彩原则 + +> 大面积冷暗，小面积暖光。 +> 泥是现实，星是精神，火是同行。 + +不要做成： + +* 金色励志大片； +* 旅游风景片； +* 商业橙蓝大片； +* 赛博蓝紫科技感； +* 过度唯美的日落封面。 + +--- + +## 6. 人物处理规则 + +封面人物必须遵守： + +* 不露脸； +* 不居中大特写； +* 不做英雄姿态； +* 不张开双臂； +* 不回头看镜头； +* 不像旅行博主； +* 不像摇滚明星海报。 + +人物应是： + +> 环境中的一个行者锚点。 + +最好是： + +* 背影； +* 稍偏画面一侧； +* 正在走； +* 外套下摆和背包轮廓清楚； +* 身体有疲惫感，但不塌。 + +--- + +## 7. 封面信息层级 + +封面第一眼看到： + +> 夜路 / 泥水 / 星光 / 远方火。 + +第二眼看到： + +> 有一个人还在走。 + +第三眼看到： + +> 歌名《在路上》。 + +不要让第一眼变成： + +> 一个中年男人的背影照。 + +--- + +## 8. 可选封面方案 + +### 方案 A：主推版 + +#### **泥水星空 + 小人物 + 远方火光** + +适合做正式 MV 封面。 + +优势： +最完整地压缩全片精神。 + +风险： +泥水星空必须处理好，否则会脏或不够诗性。 + +--- + +### 方案 B：风景长路版 + +#### **清晨泥泞长路 + 小人物背影** + +适合作为更克制的备用封面。 + +优势： +构图稳，风景叙事感强。 + +风险： +记忆点可能不如 A，容易只是“公路片封面”。 + +--- + +### 方案 C：夜路篝火版 + +#### **夜色长路 + 小火 + 远方火点** + +适合作为情绪更温暖的备用封面。 + +优势： +结尾感强，同行者意象明显。 + +风险： +容易偏露营，必须压住火的热闹感。 + +--- + +## 9. 我建议的最终封面方向 + +优先做： + +> **方案 A：泥水星空 + 小人物 + 远方火光** + +因为它最像《在路上》的精神浓缩版。 + +一句话定义： + +> **一个人走在泥泞夜路上，脚下是破碎星空，远方有微弱火光。** + +这张封面如果成立，观众不用看歌词，也能感到： + +> 这不是胜利的歌。 +> 这是继续走的歌。 + +--- + +## 10. 封面验收标准 + +封面完成后，强哥可以按这 8 条审： + +1. 第一眼是不是看到路和世界，而不是脸？ +2. 人物是否足够小？ +3. 泥水星空是否明确？ +4. 远方火光是否克制？ +5. 有没有避免露营感？ +6. 有没有避免人物写真感？ +7. 歌名字体是否克制、有风尘感？ +8. 缩小成视频缩略图后，是否仍能看出“路、火、人、星”？ + +--- + +## 11. 封面一句话标准 + +最终封面应该守住这一句： + +> **泥泞是真的，星空也是真的；人很小，但还在走，远方还有火。** diff --git a/projects/2026-06-23-在路上/project.md b/projects/2026-06-23-在路上/project.md index 3ac2247..da581d7 100644 --- a/projects/2026-06-23-在路上/project.md +++ b/projects/2026-06-23-在路上/project.md @@ -2,11 +2,13 @@ ## Status -- Status: execution skeleton initialized from cleaned intake. +- Status: shot-image production complete; assembly / closeout pending. - Medium type: 16:9 MV image/video production package. -- Current phase: execution-prep asset tests, before full 51-shot production. +- Lifecycle phase: assembly-closeout. +- Workflow phase: production-light. +- Current phase: MV shot-image production complete. - Planning state: first GPT planning round accepted and closed. -- Current execution focus: R6 Phase 2 retry callback completed; `S-20/S-22/S-23/S-24` accepted by user. +- Current execution focus: no remaining not-started shot images; next work is assembly / export / closeout when requested. ## Accepted Intake @@ -41,13 +43,17 @@ The emotional center is not victory. It is continuing. | C: 泥泞里的星空 | accepted repaired sample | Soul image; dark night/pre-dawn muddy puddle with broken starlight. | | D: 夜路篝火型 | accepted | Ending warmth and distant fellow travelers. | -## Current Assets To Materialize +## Production Base Status -1. Character anchor board. -2. Transparent character slices for execution reference. -3. Style anchor board with A/B/C/D samples. -4. C-anchor retest: dark muddy puddle with visible broken star reflection. -5. Five-shot test package before full production. +1. Character anchor board: accepted. +2. Transparent character slices for execution reference: accepted. +3. Style anchor board with A/B/C/D samples: accepted. +4. C-anchor retest: dark muddy puddle with visible broken star reflection: accepted. +5. Five-shot test package before full production: accepted. +6. Production-light route: accepted after `session-12` and `session-13`. +7. Five-slide production capacity: accepted after `S-32/S-34/S-35/S-36/S-37`; it remains historical capacity evidence. +8. Eight-slide production capacity: accepted after `S-38/S-39/S-40/S-41/S-42/S-43/S-44/S-45`; default ordinary production package size is now up to 8 slides. +9. Final production package: `S-48/S-49/S-50` completed through direct child production-light and recorded as production-pass. ## Key Local Execution Files @@ -60,9 +66,10 @@ The emotional center is not victory. It is continuing. ## Next Action -Wait for the next user production decision after R6 Phase 2 retry outputs `s20-v3`, `s22-v3`, `s23-v3`, and `s24-v2` were accepted. +Shot-image production is complete for the current 51-shot table. Do not dispatch more production image packages unless the user requests append-only repair/retry for named slides. The next normal stage is assembly / export / closeout planning. ```text +handoff/2026-06-25-on-the-road-formal-production-handoff.md handoff/2026-06-24-character-consistency-r6-notification-s16-handoff.md docs/workflows/slide-task-orchestration.md docs/workflows/slide-task-templates.md @@ -81,3 +88,7 @@ R5 brief-ownership result: `session-06` used clean child thread `019ef907-16b7-7 Package-size trials continued after R5. `session-07` completed a direct 3-slide package (`S-11`, `S-12`, `S-13`) with user-observed context 157K and elapsed time 30m5s; user accepted all three images. `session-08` completed a user-authorized 4-slide package (`S-15`, `S-16`, `S-17`, `S-18`) with completion gate pass, but user rejected `s16-v2` because it did not achieve `横向大远景，路从画面一侧切向远方，人物极小` and was too similar to `s15`. `session-09` then completed R6 Phase 1 for append-only `s16-v3` and woke the parent thread by callback rather than parent polling. R6 Phase 2 retry outputs from `session-11` for `S-20/S-22/S-23/S-24` are accepted. `S-19/s19-v2` is also accepted and should not be regenerated. The accepted quality route is `gpt-image-2` Advisor prompt/spec plus active style master, active character registry when needed, and visible reference images loaded before generation. `session-11` kept MiniMax MCP/external vision off and treated partially written v2 files as historical evidence only. + +R7 production-light simplification is accepted. `session-12` generated `S-26/S-27/S-28` as a production-light package with review `none`; user accepted all three images and observed 25m48s / 134K tokens. `session-13` generated `S-29/S-30/S-31` as a direct child session without Agent/SubAgent; user accepted all three images and observed 10m32s / 114K tokens. The production capacity validation session `019efc24-96a4-7582-bb05-914b82a726d5` generated `S-32/S-34/S-35/S-36/S-37`; user accepted all five images and reported callback at about 152K tokens / 58% context, with final observed usage about 164K tokens / 63% context. The 8-slide compression validation session `019efc48-d4af-7972-8646-6d04778dfb2c` generated `S-38/S-39/S-40/S-41/S-42/S-43/S-44/S-45`; user promoted the package to production-pass by production-stage default and reported 174K tokens / 67% context / 38m37s elapsed. The final production session `019efc78-e0f1-7601-aaec-bab4f3e6f99d` generated `S-48/S-49/S-50` and reported `complete`, `problems: none`, `needs_user: none`; parent verification found expected files present, review files absent, and no post-callback child writes. This project has completed shot-image production for the current 51-shot table. + +Capacity validation mechanism note: the child session sent callback before its final handoff update and narrow closeout checks. Future production prompts must require callback as the child session's final action. diff --git a/projects/2026-06-23-在路上/slides/s00/s00-v1-brief.md b/projects/2026-06-23-在路上/slides/s00/s00-v1-brief.md new file mode 100644 index 0000000..ac4ba23 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s00/s00-v1-brief.md @@ -0,0 +1,94 @@ +# S-00 Slide Brief V1 + +## Status + +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Version resolved by folder scan: `s00-v1` +- Design gate: passed by user +- Output status target: production-pass + +## Source Basis + +- `slides/s00/s00-v1-design.md`: approved Chinese cover design and post-production guide. +- `slides/s00/s00-v1-design-en.md`: English companion design draft. +- `intake-repair/fix-title-2026-06-25-cover-design/阶段五-封面设计骨架.fixed.md`: repaired GPT cover skeleton used as concept source. +- `slides/slides.md`: production state table to be updated after generation. +- `visual-system/style/style-prompt-master.md` +- `visual-system/characters/character-reference-registry.md` + +## Slide Function + +`S-00` is the textless MV cover base. It must carry the whole film's visual promise before the first story shot: muddy road, broken starlight, a small back-view traveler, and one restrained distant fire. It is not a normal narrative frame and not a finished text poster. + +## First Read + +Dark wet muddy road and puddle first; broken cold star reflection second; tiny traveler walking away third; distant fire last. The upper dark negative space must remain usable for manual post-production title placement. + +## Risk Class + +cover image, mixed C/D motif, distant character identity, post-production safe-area requirement + +## Reference Plan + +- Style module: mixed C / Muddy Starlight and D / Night Road Fire, with the global road-film kernel. +- Active style references: + - `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png` + - `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png` + - `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png` +- Accepted prior shot references for continuity only, not as direct copy targets: + - `projects/2026-06-23-在路上/slides/s46/s46-v1-image.png` + - `projects/2026-06-23-在路上/slides/s51/s51-v1-image.png` +- Character references: + - `projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-codex-posture-test.png` + - `projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-back-reference-crop-tight.png` + - `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png` +- Visible-context roles for generation: + - Image 1 = master character anchor for outfit, backpack, and identity continuity. + - Image 2 = walking-back crop for posture and back-view silhouette. + - Image 3 = repaired C muddy starlight reference for puddle texture and broken star reflection. + - Image 4 = D night-road fire reference for restrained distant warm signal. + +## Visual Requirements + +- 16:9 horizontal photorealistic cinematic documentary still. +- Textless base image only: no generated title, subtitle, logo, watermark, or readable sign. +- Foreground: real wet muddy road and irregular puddle in the lower 25%-35% of the frame. +- Puddle: broken cold-white star reflection visible inside dark dirty water, fragmented by mud, ripples, stones, and road grit. +- Midground: one small 50-year-old male traveler walking away or three-quarter away, dark long coat, old backpack, old boots. +- Traveler scale: about 1/10 to 1/8 of frame height, smaller than the road and world. +- Road: starts near the foreground and recedes toward the distant fire. +- Distance: one tiny restrained warm-orange roadside fire, with at most one or two weaker far fire points. +- Upper area: dark, quiet, visually clean negative space suitable for later manual title placement. + +## Character Handling + +Use the current active character anchor and walking-back crop/slice. Identity should read through silhouette, long dark coat, backpack shape, walking direction, and road-worn posture. Do not show a face or make the traveler a portrait subject. + +## Composition / Scale + +Wide road-film cover frame. Keep environment first, road second, puddle/starlight third, traveler fourth, fire fifth. The traveler should sit below or near the middle third, small enough to remain a coordinate in the landscape. The fire should be small and distant, not a camp scene and not the destination. + +## Light / Color / Texture + +Low-saturation deep blue-black night, wet dark blue-gray, earth brown mud, coal black coat, cold-white reflected stars, and one restrained warm-orange fire signal. Mud, water, stones, road grit, coat fabric, backpack wear, and boot edges should feel physically real. + +## Forbidden Drift + +No generated text, Chinese characters, English characters, subtitle, title, logo, watermark, UI, readable sign, poster layout, hero backshot, centered large person, face close-up, travel advertisement, camping lifestyle, tents, chairs, cookware, group reunion, victory, arrival, fantasy magic glow, cyberpunk, neon, bright sunrise, bright sunset, clean tourist road, clean mirror lake, or full S-46/S-51 duplication. + +## Acceptance Checklist + +- Image is textless and usable as a cover base for later manual title design. +- First read is muddy night road and puddle, not a large person or campfire. +- Broken star reflection is visible but not a full close-up repetition of `S-46`. +- Traveler is small, walking away, and anchored to the active character references. +- Distant fire is tiny and restrained, not a campsite or final reunion. +- Upper dark negative space remains available for title placement. +- Output may be recorded as production-pass under the production-stage default. + +## Prompt Handoff + +Build the final prompt through the `gpt-image-2` Advisor / host-native workflow. Load the active character anchor and walking-back crop/slice before generation, plus the C muddy starlight and D night-road fire style references. Save selected host output as `projects/2026-06-23-在路上/slides/s00/s00-v1-image.png` and update `slides/slides.md`. diff --git a/projects/2026-06-23-在路上/slides/s00/s00-v1-design-en.md b/projects/2026-06-23-在路上/slides/s00/s00-v1-design-en.md new file mode 100644 index 0000000..3df812f --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s00/s00-v1-design-en.md @@ -0,0 +1,210 @@ +# S-00 Slide Design V1 + +## Status + +- Project: `2026-06-23-在路上` +- Slide: `s00` +- Version: `v1` +- Artifact type: cover design / human-facing post-production guide +- Status: draft for user confirmation +- Design line: MV cover version +- Image generation status: not started +- Next gate: user confirms this design before `brief`, `prompt`, and image generation. + +## Design Purpose + +`s00` is the cover slide for the MV. It is different from ordinary production shots because it must support later human post-production: title placement, typography, color, and possible platform thumbnail use. + +Therefore this file is for the user and post-production decisions. The later `brief.md` remains the execution card for Codex image generation, and the later `prompt.md` remains the image-model prompt. The generated image should be a clean textless base image. + +## Discussion Summary + +The accepted direction is the more MV-cover-like version: keep a small back-view traveler and distant fire, while preserving the project's core rule that the world, road, mud, night, and starlight are read before the person. + +The GPT cover skeleton proposed "muddy road, broken starlight in puddle, distant fire, small traveler." After looking at the already accepted project images and visual system, the direction was tightened: + +- Do not make the cover a literal collage of `S-46` and `S-51`. +- Do not reveal the whole ending too early. +- Do not make the traveler or fire too prominent. +- Do not let the image contain generated title text. +- Use `design` for human/post-production information and keep `brief` compact for execution. + +The final cover should feel like a distilled MV key image, not a poster with all answers explained. + +## Visual Concept + +Working concept: + +> 泥路低处见星，远方有火，人仍在走。 + +The image is a dark pre-dawn or night road-film still. The viewer first reads wet mud, a road, a low puddle with broken cold starlight, and darkness. The second read finds a small back-view traveler continuing forward. The third read notices a restrained orange fire far ahead. + +The emotional meaning is: + +> 泥泞是真的，星空也是真的；人很小，但还在走，远方还有火。 + +This should not feel triumphant. It should feel quiet, road-worn, unfinished, and still moving. + +## Image Base Requirements + +The generated image must be a 16:9 horizontal textless base image. + +Core image structure: + +- Foreground: muddy wet road and a real puddle occupying roughly the lower 25%-35% of the frame. +- Puddle: visible broken cold-white starlight reflected in dirty water, but less close and less complete than the accepted `S-46` soul shot. +- Midground: one small 50-year-old male traveler, back view or side-back view, dark long coat, old backpack, old boots, walking away. +- Traveler scale: about 1/10 to 1/8 of frame height, small enough to stay secondary. +- Road: wet muddy track leads from foreground toward distance; road direction should guide the eye toward the far fire. +- Distance: one tiny restrained orange roadside fire plus at most one or two weaker fire points farther away. +- Sky / upper area: keep enough dark, quiet negative space for later title placement. +- Overall style: photorealistic cinematic documentary road-film still, restrained poetic realism, low saturation. + +The generated image must not contain: + +- title text; +- subtitles; +- labels; +- logo; +- watermark; +- readable signs; +- poster typography; +- AI-generated Chinese or English characters. + +## Text / Typography Plan + +Default title text for post-production: + +```text +在路上 +``` + +Do not render this title into the generated image. Add it later in post-production. + +Default subtitle: none. + +Optional subtitle only if the thumbnail feels too quiet after image review: + +```text +泥泞里，有星空 +``` + +Typography direction: + +- Chinese title should feel road-worn, restrained, and human. +- Prefer a rough but legible hand-brushed title style, or a thick Song/Hei hybrid with slight weathered texture. +- Avoid modern technology fonts, glossy commercial poster fonts, elegant literary thin fonts, and over-dramatic calligraphy. +- Keep the title simple; do not add decorative frames or slogan blocks. + +Suggested title color: + +- Main title: warm off-white / aged paper, approximately `#F2E6CF`. +- Optional subtitle: muted warm gray, approximately `#BFAF96`. +- Shadow or stroke: very dark blue-black, approximately `#071016`, used lightly only for readability. + +Optional accent: + +- A very restrained warm-orange echo from the distant fire, approximately `#C67935`, only for a tiny underline, dot, or metadata accent if needed. Do not make the title orange. + +## Layout / Safe Areas + +Default title placement: + +- Upper left or upper right, depending on where the generated sky/negative space is cleanest. +- Keep at least 7%-9% margin from the frame edge. +- Title block should not cover the puddle, traveler, road direction, or far fire. + +Preferred layout if the generated image follows the intended composition: + +- Title in upper-left dark sky area. +- No subtitle. +- Keep all text above the horizon/road action zone. +- Let the lower third remain purely visual: mud, puddle, broken starlight, and road. + +Fallback layout: + +- If the upper-left area has too much visual activity, move title to upper-right. +- If both upper corners are active, use a small title block in the darker side of the sky, not on top of the puddle. + +Thumbnail rule: + +- At small size, viewers should still read: road / fire / small person / night mood. +- The title should remain legible but not dominate the image. + +## Color Guidance + +Image palette: + +- deep blue-black; +- wet dark blue-gray; +- cold gray; +- earth brown; +- dark khaki; +- coal black. + +Emotional highlights: + +- cold-white broken starlight in the puddle; +- small warm-orange fire in the distance; +- faint dark pre-dawn blue if needed. + +Color balance: + +> 大面积冷暗，小面积暖光。泥是现实，星是精神，火是同行。 + +Avoid: + +- gold inspirational poster tone; +- orange-blue blockbuster grading; +- purple/blue sci-fi tone; +- tourist sunrise or sunset; +- clean glossy commercial contrast. + +## Post-Production Notes + +The image is expected to be a base image for manual title work. + +Post-production steps after a usable image exists: + +1. Add the title `在路上` manually. +2. Choose upper-left or upper-right placement based on the actual generated negative space. +3. Use warm off-white title color with light dark shadow/stroke only if readability needs it. +4. Keep subtitle disabled by default. +5. If a subtitle is needed for platform thumbnail readability, use `泥泞里，有星空` in small muted warm gray. +6. Do not add explanatory copy, credit blocks, decorative frames, or extra graphic elements unless a later task explicitly asks for a poster variant. + +## Constraints / Avoid + +Avoid image drift: + +- no person-centered poster; +- no large face or front-facing portrait; +- no heroic pose; +- no arms-open silhouette; +- no traveler-blog backshot; +- no camping scene; +- no group reunion; +- no victory feeling; +- no clean tourism road; +- no bright sunrise/sunset; +- no fantasy portal or magic glow; +- no sci-fi/cyberpunk palette; +- no text in image. + +Avoid cover-specific mistakes: + +- do not make the puddle so close that it becomes only `S-46` again; +- do not make the fire so clear that it becomes the ending `S-51`; +- do not make the traveler larger than the world; +- do not fill all dark sky, because post-production needs title space. + +## Confirmation Gate + +If the user confirms this design, the next step is: + +1. Write `s00-v1-brief.md`. +2. Write `s00-v1-prompt.md` through the required `gpt-image-2` Advisor / host-native workflow. +3. Generate `s00-v1-image.png` as a textless base image. +4. Update `projects/2026-06-23-在路上/slides/slides.md` with `s00` progress. +5. Record the generated image as `production-pass` by default because the project is in production-light and this task's images are user-authorized as pass-by-default unless the user later requests an append-only retry. + diff --git a/projects/2026-06-23-在路上/slides/s00/s00-v1-design.md b/projects/2026-06-23-在路上/slides/s00/s00-v1-design.md new file mode 100644 index 0000000..c6edf6a --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s00/s00-v1-design.md @@ -0,0 +1,212 @@ +# S-00 封面设计稿 V1 + +## 状态 + +- 项目：`2026-06-23-在路上` +- Slide：`s00` +- 版本：`v1` +- 文档类型：封面设计稿 / 人类后期制作说明 +- 当前状态：待用户确认 +- 设计方向：MV 封面版 +- 生图状态：未开始 +- 下一步确认门：用户确认本设计稿后，再进入 `brief -> prompt -> image`。 + +## 设计目的 + +`s00` 是《在路上》MV 的封面页，不是普通正文镜头。正文镜头基本只需要 `brief -> prompt -> image`，但封面还要服务后期制作，包括标题摆放、字体气质、颜色、留白和平台缩略图可读性。 + +因此本文件给用户和后期制作使用。后续 `brief.md` 只保留 Codex 生图执行所需信息，`prompt.md` 只保留模型提示词。封面图片本身必须是无文字底图，标题在后期手动添加。 + +英文说明稿保存在同目录： + +```text +s00-v1-design-en.md +``` + +## 本次讨论总结 + +GPT 给出的封面骨架方向是“泥路尽头，泥水有星，远方有火”：一条泥泞长路、前景泥水星空、小人物背影、远处火光。这个方向与项目视觉系统一致，但 GPT 当时看不到已经生成并接受的 `S-01`、`S-46`、`S-51` 等实际画面。 + +结合已完成镜头后的修正是： + +- 不把封面做成 `S-46` 泥水星空近景和 `S-51` 远火结尾的简单拼贴。 +- 不提前把全片的答案都讲满。 +- 保留小背影和远火，让它更像 MV 封面，而不是纯开场空镜。 +- 画面不能出现 AI 生成文字，所有标题、字幕、标识都留给后期。 +- `design` 记录人类设计与后期制作要求；`brief` 保持简洁，专门给 Codex 执行。 + +最终选择：更 MV 封面版。 + +## 核心概念 + +工作概念： + +> 泥路低处见星，远方有火，人仍在走。 + +封面第一眼应该看到夜路、泥水、星光和冷暗世界。第二眼看到一个很小的行者背影仍在往前走。第三眼看到远方有一簇很克制的火。 + +情绪不是胜利，不是抵达，也不是会师，而是： + +> 泥泞是真的，星空也是真的；人很小，但还在走，远方还有火。 + +这张封面应该像整支 MV 的钥匙，而不是剧情结论。 + +## 底图生成要求 + +生成图必须是 16:9 横版、无文字底图。 + +画面结构： + +- 前景：湿泥路和真实泥水洼，占画面下方约 25%-35%。 +- 水洼：能看到破碎冷白星光倒影，但不要像 `S-46` 那样完整进入灵魂近景。 +- 中景：一个 50 岁左右男性行者的小背影或侧背影，深色长外套、旧背包、旧靴，继续向远处走。 +- 人物比例：约为画面高度的 1/10 到 1/8，必须小于世界和道路。 +- 道路：湿泥路从前景延伸到远方，引导视线通向火光。 +- 远景：一簇很小的暖橙色路边火，最多再有一两个更弱的远方火点。 +- 天空 / 上方区域：保留足够干净的暗部负空间，方便后期放标题。 +- 风格：现实纪实质感 + 克制诗性现实主义 + 公路纪录片式风景叙事。 + +生成图禁止出现： + +- 标题文字； +- 副标题； +- 任何中文或英文字符； +- 标牌可读文字； +- logo； +- 水印； +- 海报排版元素。 + +## 文字与字体方案 + +默认后期标题： + +```text +在路上 +``` + +默认不放副标题。 + +备用副标题，仅当平台缩略图需要更明确表达时使用： + +```text +泥泞里，有星空 +``` + +字体气质： + +- 有路上风尘感； +- 粗粝但清楚； +- 温暖、克制、有人味； +- 可偏手写、手刷、粗宋黑混合或旧感标题字； +- 不要科技感、商业大片感、花体文艺感、过度书法感。 + +文字颜色建议： + +- 主标题：暖白 / 旧纸色，建议 `#F2E6CF`。 +- 备用副标题：低饱和暖灰，建议 `#BFAF96`。 +- 阴影或描边：深蓝黑，建议 `#071016`，只在可读性不足时轻微使用。 +- 点缀色：可极少使用火光橙，建议 `#C67935`，但标题本身不要做成橙色。 + +## 版式与安全区 + +默认标题位置： + +- 画面左上或右上，根据实际底图哪一侧暗部更干净决定。 +- 距离画面边缘至少保留 7%-9% 安全边距。 +- 标题不要压住泥水星空、人物、道路方向或远方火光。 + +优先版式： + +- 标题放在左上暗天空区域。 +- 不放副标题。 +- 下三分之一完全留给泥、水、星光和路。 +- 人物和远火保持在视觉叙事里，不被文字遮挡。 + +备选版式： + +- 如果左上天空太乱，标题移到右上。 +- 如果两侧天空都复杂，标题放在更暗、更空的一侧，但不要贴近水洼。 +- 若加副标题，副标题放在主标题下方，字号显著小于主标题。 + +缩略图规则： + +- 小尺寸下仍要能读到：路、火、小人、夜色。 +- 标题要清楚，但不能变成整张图的第一主角。 + +## 色彩系统 + +底图主色： + +- 深蓝黑； +- 湿暗蓝灰； +- 冷灰； +- 泥土褐； +- 暗卡其； +- 煤黑。 + +情绪色： + +- 泥水里的冷白星光； +- 远处火光的极小暖橙； +- 必要时有一点黎明前暗蓝。 + +色彩原则： + +> 大面积冷暗，小面积暖光。泥是现实，星是精神，火是同行。 + +避免： + +- 金色励志大片； +- 橙蓝商业大片； +- 赛博蓝紫科技感； +- 旅游日出 / 日落； +- 露营广告暖色； +- 过度唯美、过度干净的风景封面。 + +## 后期制作说明 + +底图生成后，后期步骤建议： + +1. 手动添加标题 `在路上`。 +2. 根据底图负空间选择左上或右上。 +3. 使用暖白 / 旧纸色标题，必要时加极轻的深色阴影或描边。 +4. 默认不加副标题。 +5. 如果缩略图信息不足，再加入小号副标题 `泥泞里，有星空`。 +6. 不加入解释性文案、制作说明、装饰框、过多图形元素。 +7. 不把封面做成“海报设计展示”，保持 MV 封面底图的电影静帧感。 + +## 禁止偏移 + +画面层面禁止： + +- 人物居中大背影； +- 正脸或半脸大特写； +- 英雄姿态； +- 张开双臂； +- 旅行博主式背影照； +- 露营场景； +- 人群会师； +- 胜利、抵达、团圆； +- 干净旅游公路； +- 明亮日出日落； +- 奇幻魔法光效； +- 科技 / 赛博色调； +- 任何文字出现在图中。 + +封面层面特别注意： + +- 水洼不能太近，否则会变成 `S-46` 的重复。 +- 火光不能太明确，否则会提前变成 `S-51` 的结尾。 +- 人不能大过路和世界。 +- 天空和暗部不能被填满，因为后期需要放标题。 + +## 确认门 + +用户确认本设计稿后，下一步执行： + +1. 写 `s00-v1-brief.md`。 +2. 按 `gpt-image-2` Advisor / host-native 路线写 `s00-v1-prompt.md`。 +3. 生成无文字底图 `s00-v1-image.png`。 +4. 更新 `projects/2026-06-23-在路上/slides/slides.md` 的 `s00` 进度。 +5. 由于项目处于生产阶段，本次生成图默认按 `production-pass` 记录；如后续效果不可用，再走 append-only retry。 + diff --git a/projects/2026-06-23-在路上/slides/s00/s00-v1-image.png b/projects/2026-06-23-在路上/slides/s00/s00-v1-image.png new file mode 100644 index 0000000..e72ac73 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s00/s00-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s00/s00-v1-prompt.md b/projects/2026-06-23-在路上/slides/s00/s00-v1-prompt.md new file mode 100644 index 0000000..36c2f72 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s00/s00-v1-prompt.md @@ -0,0 +1,121 @@ +# S-00 Slide Prompt V1 + +## Status + +- Used once for host-native image generation. +- Review mode: none. +- Target slide status after generation: production-pass. +- Generated image: `s00-v1-image.png`. +- Design gate: passed by user. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s00/s00-v1-design.md` +- `projects/2026-06-23-在路上/slides/s00/s00-v1-brief.md` + +## Reference Inputs + +- Style module: mixed C / Muddy Starlight and D / Night Road Fire, with the global road-film kernel. +- Active style references: + - `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png` + - `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png` + - `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png` +- Character references: + - `projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-codex-posture-test.png` + - `projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-back-reference-crop-tight.png` + - `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png` +- Accepted prior shot references for continuity only: + - `projects/2026-06-23-在路上/slides/s46/s46-v1-image.png` + - `projects/2026-06-23-在路上/slides/s51/s51-v1-image.png` +- Generation visible-context roles: + - Image 1 = master character anchor for identity, clothing, backpack, and posture continuity. + - Image 2 = walking-back character crop for back-view stride and silhouette. + - Image 3 = repaired C muddy starlight reference for real mud, dark water, and broken star reflection. + - Image 4 = D night-road fire reference for small restrained warm fire in darkness. + +## Operation Notes + +- Project-bound prompt produced through `gpt-image-2` Skill Advisor / host-native workflow. +- Garden/API expected: no. +- Actual image references must be loaded into visible context before host generation. +- The generated image must contain no text because title and subtitle are handled manually in post-production per `s00-v1-design.md`. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host has built-in `image_gen`. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Source files read: + - `AGENTS.md` + - `VIDEO_WORKBENCH.md` + - `docs/workflows/slide-task-orchestration.md` + - `docs/workflows/slide-task-templates.md` + - `projects/2026-06-23-在路上/slides/s00/s00-v1-design.md` + - `projects/2026-06-23-在路上/slides/s00/s00-v1-design-en.md` + - `projects/2026-06-23-在路上/slides/slides.md` + - `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` + - `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + +## Final Prompt + +Use case: photorealistic-natural + +Asset type: 16:9 textless MV cover base image for `在路上`, slide `S-00`. + +Input images: Image 1 is the master character anchor for identity, clothing, backpack, and posture continuity. Image 2 is the walking-back character crop for the traveler's back-view stride and silhouette. Image 3 is the repaired muddy starlight reference for dark wet mud, real puddle texture, and broken cold-white star reflection. Image 4 is the night-road fire reference for one tiny restrained warm roadside fire in deep darkness. + +Primary request: + +Create one 16:9 horizontal photorealistic cinematic documentary road-film still for the textless MV cover of `在路上`. The image should say: low in the muddy road there are stars, far away there is a fire, and the small traveler is still walking. It must be a cover base with no text of any kind. + +Scene/backdrop: + +A dark wet wilderness road at night or deep pre-dawn. The ground is muddy, uneven, physically real, and wet, with old ruts, stones, road grit, puddle edges, and dirty water. The road begins in the lower foreground and recedes into a dark distance. The upper half of the frame contains quiet dark sky or dark negative space, clean enough for a manual title to be added later. + +Foreground: + +In the lower 25% to 35% of the frame, show real wet mud and an irregular muddy puddle. Inside the puddle, show a broken cold-white star reflection: small star points fragmented by ripples, mud, stones, and dirty water texture. The puddle should be readable but not an extreme close-up and not a direct repetition of the accepted `S-46` soul shot. + +Subject: + +A small 50-year-old male traveler walking away or three-quarter away along the muddy road. Use Image 1 and Image 2 to preserve the current character identity through silhouette: long dark road-worn coat, old backpack, old boots, steady upright walking posture, tired but still capable. The traveler is not a portrait subject, has no visible face, and should occupy only about 1/10 to 1/8 of frame height. + +Distance and fire: + +Place one tiny restrained warm-orange roadside fire far ahead along the road, with at most one or two much weaker distant fire points. The fire is a small signal in darkness, not a destination, not a camp scene, and not a reunion. + +Composition/framing: + +Environment first, road second, muddy puddle with broken stars third, small back-view traveler fourth, distant fire last. Use a wide road-film cover composition. Keep the world larger than the person and the road longer than the person. Preserve enough clean dark upper negative space for later manual title placement. The title area must stay empty in the image itself. + +Lighting/mood: + +Low-saturation deep blue-black night, wet dark blue-gray, earth brown mud, coal black coat, cold-white reflected stars, and one restrained warm-orange fire signal. Large cold darkness, very small warmth. Mood is unfinished, restrained, road-worn, and quietly persistent: not victory, not arrival, not despair. + +Texture and realism: + +Photorealistic documentary texture with restrained poetic realism. Real wet mud, real dirty puddle water, stones, grit, boot marks, damp coat fabric, old backpack fabric, worn boot edges, subtle haze or moisture in air. No glossy fantasy effect. + +Strict constraints: + +No generated text anywhere. No title, subtitle, Chinese characters, English characters, readable sign, logo, watermark, UI, label, poster layout, frame graphic, or typography. This is a textless base image for later manual post-production. + +## Negative / Avoid + +Avoid travel advertising, scenic tourism, motivational poster, heroic centered backshot, large person, face close-up, portrait MV cover, fashion photography, clean tourist road, bright sunrise, bright sunset, orange-blue blockbuster look, cyberpunk, neon, sci-fi UI, code, fantasy magic glow, glowing portal, clean mirror lake, full-body hero pose, arms open, camping lifestyle, tent, camping chair, cookware, picnic setup, cozy campsite, group gathering, reunion, victory, arrival, large theatrical bonfire, text, letters, numbers, labels, logos, watermarks, UI. + +## Output Intent + +Save selected host output as `projects/2026-06-23-在路上/slides/s00/s00-v1-image.png` and record `s00` as production-pass in `projects/2026-06-23-在路上/slides/slides.md`. + +## Generation Result + +- Host default output: `C:\Users\wangq\.codex\generated_images\019efc4e-7791-7be1-940f-ce1b7318fea0\ig_0c84b336dc89f00a016a3c8c4d4bb4819185eab6eb16919462.png` +- Project copy: `projects/2026-06-23-在路上/slides/s00/s00-v1-image.png` +- Image size: `1672 x 941` +- SHA256: `85164F5385BD0D001DB63C3E5F92847000FFC7168D330E7372D9CDBD0A45F4C7` +- Review file: none; production-stage default records this image as production-pass. diff --git a/projects/2026-06-23-在路上/slides/s26/s26-v2-brief.md b/projects/2026-06-23-在路上/slides/s26/s26-v2-brief.md new file mode 100644 index 0000000..4212f4e --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s26/s26-v2-brief.md @@ -0,0 +1,69 @@ +# S-26 Slide Brief V2 + +## Status + +- Status: ready for one-shot slide agent. +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Overwrite allowed: no. + +## Source Basis + +- `slides/slides.md`: Night water/starlight begins but is not full reveal. +- `style-prompt-master.md`: Module C, with restraint from the accepted repaired C sample. +- Existing broad-run `s26-v1-image.png` is historical evidence only. + +## Slide Function + +Start the C sequence without using the full S-46 soul-shot reveal. The frame should let night water, mud, and very faint broken starlight begin to appear. + +## First Read + +Dark wet mud and water at night; only after that, faint broken star points beginning in the water. + +## Risk Class + +core motif; environment-led; no visible identity required. + +## Reference Plan + +- Style module: C muddy starlight. +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png`; role: C repaired muddy-water / broken-starlight style reference. +- Image 2: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; role: overall low-saturation documentary road-film calibration. +- Character reference: none by default; do not introduce a full traveler. + +## Visual Requirements + +- 16:9 horizontal photorealistic cinematic documentary still. +- Real mud, dirty water, uneven puddle edge, grit, stones, weak ripples. +- Star reflection is early and partial, not a full spectacular reveal. +- Keep darkness physical and low; no pretty mirror lake. + +## Character Handling + +No visible character is required. If a traveler trace appears, keep it below 10% as an edge trace only, with no face or full body. + +## Composition / Scale + +Low ground-level or shallow roadside view. Mud and water dominate; the reflection is discovered inside dirty water rather than displayed as sky scenery. + +## Light / Color / Texture + +Cold blue-gray, wet black-gray, mud brown, coal black, very small cold-white star points. No sunrise, no sunset, no warm horizon. + +## Forbidden Drift + +No cosmic spectacle, magic portal, shoe ad, clean lake, sunrise, full-body traveler, portrait, large real sky, readable text, logos, UI, watermark. + +## Acceptance Checklist + +- Mud/water reads before starlight. +- Starlight is faint, broken, and physically interrupted. +- The shot feels like C beginning, not S-46 final reveal. +- Review file remains absent. + +## Prompt Handoff + +Carry Image 1 and Image 2 as visible-context references when possible. The final prompt must record `gpt-image-2` Advisor / host-native mode, exact source files read, and target `s26-v2-image.png`. + diff --git a/projects/2026-06-23-在路上/slides/s26/s26-v2-image.png b/projects/2026-06-23-在路上/slides/s26/s26-v2-image.png new file mode 100644 index 0000000..40101f6 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s26/s26-v2-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s26/s26-v2-prompt.md b/projects/2026-06-23-在路上/slides/s26/s26-v2-prompt.md new file mode 100644 index 0000000..fbf897b --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s26/s26-v2-prompt.md @@ -0,0 +1,87 @@ +# S-26 Slide Prompt V2 + +## Status + +- Status: ready for host-native generation. +- Slide/version: `S-26` / `s26-v2`. +- Running mode: prompt and generate. +- Review mode: none. +- Overwrite allowed: no. + +## Based On Brief + +- Brief: `projects/2026-06-23-在路上/slides/s26/s26-v2-brief.md` +- Slide function: begin the C muddy-starlight sequence without becoming the full S-46 reveal. +- First-read order: real night mud and dirty water first; faint broken starlight discovered second. + +## Reference Inputs + +- Style module: C / muddy starlight. +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png`; role: primary C muddy-water / broken-starlight style reference; use for real wet mud, dirty puddle texture, cold faint reflected star points, and restrained ground-level framing. +- Image 2: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; role: overall style calibration; use for low-saturation documentary road-film palette and anti-drift boundaries. +- Character reference: none required; do not introduce a full traveler. +- Generation path note: Image 1 and Image 2 are visible-context references for host-native generation, not edit targets. + +## Operation Notes + +- Target image path: `projects/2026-06-23-在路上/slides/s26/s26-v2-image.png` +- Garden/API expected: no. +- Host image generation expected: yes, one image only. +- No review file should be written. + +## Advisor / Host-Native Record + +- Skill mode command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Skill mode result: + +```json +{ + "mode": "B-or-C", + "recommendation": "host-or-advisor", + "garden_mode_enabled": false, + "has_api_key": false, + "base_url": "https://api.openai.com/v1", + "model": "gpt-image-2", + "env_flag_value": "(unset)", + "summary": "MODE B / C: use host-native image generation if available; otherwise Advisor prompt only." +} +``` + +- Host-native interpretation: built-in host image generation is available, so use Mode B. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Project/source files read: + - `AGENTS.md` + - `docs/workflows/slide-task-orchestration.md` + - `docs/workflows/slide-task-templates.md` + - `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s26-v2-slide-agent-packet.md` + - `projects/2026-06-23-在路上/slides/s26/s26-v2-brief.md` + - `projects/2026-06-23-在路上/slides/slides.md` + - `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` + - `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + +## Final Prompt + +Use case: photorealistic-natural. +Asset type: 16:9 cinematic documentary road-film still for slide S-26. +Input images: Image 1 is the primary C muddy-water / broken-starlight style reference; Image 2 is the overall low-saturation documentary road-film calibration board. Use both as visible-context references, not as edit targets. + +Create a 16:9 horizontal photorealistic cinematic documentary still at night, low to the ground beside a muddy roadside puddle. The frame must first read as real wet mud, dirty water, grit, stones, broken shoe marks, uneven puddle edge, and weak ripples. Only after that should the viewer discover a few faint cold-white broken star points reflected inside the dirty water. + +Keep the starlight early, partial, and physically interrupted by mud, stones, ripples, and water texture. It must feel like the beginning of the C sequence, not the full soul-image reveal. The puddle is shallow and imperfect, not a clean mirror lake. Darkness stays physical and low, with cold blue-gray, wet black-gray, coal black, dark mud brown, and a few tiny cold-white star points. + +Composition: ground-level or very low roadside perspective. Mud and water dominate the frame. The reflected stars occupy only part of the puddle, with many dark and muddy interruptions. If any traveler trace appears, it may only be an ambiguous boot edge, trouser edge, or mud-dark shape at the extreme border, below 10% of the frame, with no face, no torso, and no full body. No readable text. + +Style: real documentary texture with restrained poetic realism, like a road-film still. Use Image 1 for the repaired C texture, mood, puddle scale, and cold broken starlight. Use Image 2 for the broader project palette and anti-drift discipline: low saturation, road-film realism, environment before person, no heroic framing. + +## Negative / Avoid + +No cosmic spectacle, no galaxy portal, no fantasy magic glow, no clean mirror lake, no sunrise, no sunset, no warm horizon, no large real sky view, no postcard landscape, no shoe advertisement, no product photography, no full-body traveler, no face, no portrait, no campfire, no city neon, no sci-fi UI, no readable text, no logos, no watermark. + +## Output Intent + +- Save the generated image as `projects/2026-06-23-在路上/slides/s26/s26-v2-image.png`. +- Keep `s26-v2-review.md` absent because review mode is none. diff --git a/projects/2026-06-23-在路上/slides/s27/s27-v2-brief.md b/projects/2026-06-23-在路上/slides/s27/s27-v2-brief.md new file mode 100644 index 0000000..ce5f606 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s27/s27-v2-brief.md @@ -0,0 +1,70 @@ +# S-27 Slide Brief V2 + +## Status + +- Status: ready for one-shot slide agent. +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Overwrite allowed: no. + +## Source Basis + +- `slides/slides.md`: Traveler stops before mud/water. +- `style-prompt-master.md`: A / C transition, landscape-led road before person. +- Existing broad-run `s27-v1-image.png` is historical evidence only. + +## Slide Function + +Show the traveler pausing before muddy water, as if the C motif is about to become legible. The person is present but still secondary to road, mud, and water. + +## First Read + +Muddy road and dark water before the traveler; traveler becomes a small coordinate at the edge of hesitation. + +## Risk Class + +medium human; motif transition. + +## Reference Plan + +- Style module: A / C mixed. +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; role: overall documentary road-film calibration. +- Image 2: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`; role: landscape-led muddy road and world-before-person composition. +- Image 3: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png`; role: muddy water / faint broken starlight cue. +- Image 4: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`; role: back-view traveler silhouette, long coat, backpack, posture. + +## Visual Requirements + +- 16:9 horizontal photorealistic cinematic documentary still. +- Traveler pauses before mud/water, not posing for a portrait. +- Long road, puddles, dark water, wet mud, stones, and weak night/pre-dawn light dominate. +- C motif is foreshadowed through water and faint broken reflection, not fully revealed. + +## Character Handling + +Use Image 4 as visible identity/silhouette reference. Keep traveler back or three-quarter back, small-to-medium-small, about 1/12 to 1/8 of the frame. No face emphasis. + +## Composition / Scale + +Environment first, road/water second, traveler third. Place the traveler near a puddle edge or road side, stopped but still oriented toward the continuing road. + +## Light / Color / Texture + +Cold gray, wet blue-gray, mud brown, dark khaki, coal black. Reflection is weak and broken. Avoid warm motivational light. + +## Forbidden Drift + +No hero backshot, no centered portrait, no clean inspirational road, no full S-46 star reveal, no fantasy glow, no tourism, no readable text, logos, UI, watermark. + +## Acceptance Checklist + +- Road and mud/water read before the person. +- Traveler identity is silhouette-stable through coat/backpack/posture. +- The pause feels unresolved and continuing, not victorious. +- Review file remains absent. + +## Prompt Handoff + +Carry Images 1-4 as visible-context references when possible. The final prompt must record `gpt-image-2` Advisor / host-native mode, exact source files read, and target `s27-v2-image.png`. + diff --git a/projects/2026-06-23-在路上/slides/s27/s27-v2-image.png b/projects/2026-06-23-在路上/slides/s27/s27-v2-image.png new file mode 100644 index 0000000..85106d8 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s27/s27-v2-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s27/s27-v2-prompt.md b/projects/2026-06-23-在路上/slides/s27/s27-v2-prompt.md new file mode 100644 index 0000000..aaaffff --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s27/s27-v2-prompt.md @@ -0,0 +1,81 @@ +# S-27 Slide Prompt V2 + +## Status + +- Status: ready for host-native generation. +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Garden/API expected: no. +- Overwrite allowed: no. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s27/s27-v2-brief.md` + +## Reference Inputs + +- Style module: mixed A / C. +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; role: overall documentary road-film calibration and avoid-drift board. +- Image 2: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`; role: landscape-led muddy-road composition, world-before-person scale, cold dawn/pre-dawn terrain. +- Image 3: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png`; role: muddy water, wet grit, faint broken starlight/reflection cue. +- Image 4: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`; role: back-view traveler silhouette continuity, long coat, backpack, posture. +- Character reference: Image 4 is required because the traveler is visible. +- Generation path note: actual reference images must be loaded in visible context before host image generation; file paths here are traceability. + +## Operation Notes + +- Target image path: `projects/2026-06-23-在路上/slides/s27/s27-v2-image.png` +- First-read order: mud / water / road before traveler. +- Scale band: traveler about 1/12 to 1/8 of the frame, back or three-quarter back, secondary to environment. +- C motif: foreshadowed through water and faint broken reflection, not fully revealed. + +## Advisor / Host-Native Record + +- `gpt-image-2` mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode check result: `mode="B-or-C"`, `recommendation="host-or-advisor"`, `garden_mode_enabled=false`, `has_api_key=false`, `model="gpt-image-2"`. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Project source files read: + - `AGENTS.md` + - `docs/workflows/slide-task-orchestration.md` + - `docs/workflows/slide-task-templates.md` + - `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s27-v2-slide-agent-packet.md` + - `projects/2026-06-23-在路上/slides/s27/s27-v2-brief.md` + - `projects/2026-06-23-在路上/slides/slides.md` + - `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` + - `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + +## Final Prompt + +Create one 16:9 horizontal photorealistic cinematic documentary road-film still for slide S-27. + +Use the attached references as follows: +Image 1 is the overall visual calibration board: real documentary texture, restrained poetic realism, low saturation, road-film composition, no travel-ad or heroic poster drift. +Image 2 is the main landscape and composition reference: a large cold muddy wilderness road where the world and road read before the person. +Image 3 is the mud/water texture and faint broken-reflection reference: real wet mud, grit, stones, dark water, weak cold-white starlight fragments in a puddle. +Image 4 is the required traveler identity/silhouette reference: back-view road-worn traveler with long dark coat, backpack, tired upright posture, no face emphasis. + +Scene: a cold pre-dawn or deep blue-gray night-to-dawn muddy road. The traveler has stopped at the edge of a dark puddle or muddy water crossing, still facing the continuing road. The road, wet ruts, puddles, stones, mud, rough earth, and dark water dominate the frame. The traveler is only a small coordinate of hesitation, about 1/12 to 1/8 of the frame, back or three-quarter back, placed near the puddle edge or road side, not centered as a hero. + +Composition: environment first, road and muddy water second, traveler third. Use a medium-wide to wide road-film frame with a strong receding road line. Keep the traveler secondary and partially absorbed by the landscape. The pause should feel unresolved, as if he may continue after this moment. + +Lighting and color: cold gray, wet blue-gray, mud brown, dark khaki, coal black. Very low saturation. Weak natural pre-dawn or night ambient light only. The puddle may contain faint broken cold-white star-like reflections, fragmented by ripples, mud, stones, and dirty water. The reflection is a quiet cue, not a fantasy effect. + +Texture: real wet mud, dirty water, grit, stones, worn boot marks, road ruts, damp fabric, road-worn backpack and coat. Keep documentary realism and physical texture stronger than visual spectacle. + +Character handling: preserve the back-view silhouette, long coat, backpack, shoulder line, and walking-away posture from Image 4. Do not invent a frontal face. Do not make the traveler a fashion portrait or heroic backshot. + +Output: cinematic still, 16:9, no text, no subtitles, no logos, no UI, no watermark. + +## Negative / Avoid + +Avoid travel advertising, scenic tourism, inspirational clean road, victory tone, heroic centered stance, large portrait subject, fashion photography, broad warm sunrise, broad sunset, high saturation, fantasy glow, magic portal, cosmic spectacle, clean mirror lake, full star reveal, campfire, camping gear, readable text, labels, logos, watermarks, UI. + +## Output Intent + +- Write one production-light image to `projects/2026-06-23-在路上/slides/s27/s27-v2-image.png`. +- Do not write `s27-v2-review.md`. diff --git a/projects/2026-06-23-在路上/slides/s28/s28-v2-brief.md b/projects/2026-06-23-在路上/slides/s28/s28-v2-brief.md new file mode 100644 index 0000000..329b412 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s28/s28-v2-brief.md @@ -0,0 +1,70 @@ +# S-28 Slide Brief V2 + +## Status + +- Status: ready for one-shot slide agent. +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Overwrite allowed: no. + +## Source Basis + +- `slides/slides.md`: Bridge/round structure shadow suggests gear-like night. +- `style-prompt-master.md`: Module B, high-angle city-edge pressure. +- Existing broad-run `s28-v1-image.png` is historical evidence only. + +## Slide Function + +Show system pressure through a bridge or round structural shadow at night. The structure should suggest gear-like force without becoming sci-fi or literal machinery. + +## First Read + +Massive bridge/round structure shadow and wet city-edge ground before the traveler. + +## Risk Class + +distant human; city-pressure environment. + +## Reference Plan + +- Style module: B high-angle city edge. +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; role: overall documentary road-film calibration. +- Image 2: `projects/2026-06-23-在路上/intake/imgs/B.高位观察型蓝图.png`; role: city-edge / concrete / high-observation pressure reference. +- Image 3: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png`; role: walking side silhouette, coat, backpack, stride continuity. +- Image 4: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`; role: walking-away silhouette and backpack/coat continuity. + +## Visual Requirements + +- 16:9 horizontal photorealistic cinematic documentary still. +- Massive concrete/steel bridge, circular ramp, underpass curve, or round structural shadow. +- Gear-like feeling should come from shadow geometry and urban pressure, not from literal gears, UI, code, or sci-fi design. +- Wet ground, puddles, stains, rust traces, drain lines, hard city-edge texture. + +## Character Handling + +Use Images 3-4 as visible silhouette references. Keep the traveler small, about 1/12 to 1/8 of the frame, secondary to the built structure. No face detail. + +## Composition / Scale + +High-angle or distant observational frame. Structure and shadow dominate; traveler appears at lower side or edge as a moving coordinate. + +## Light / Color / Texture + +Cold gray concrete, wet blue-gray ground, coal black shadows, dark khaki clothing, low saturation. Night or deep pre-dawn city edge. + +## Forbidden Drift + +No cyberpunk, neon, sci-fi UI, literal machine gears, code, clean CBD promotion, glass-tower ad, fashion portrait, hero stance, readable text, logos, UI, watermark. + +## Acceptance Checklist + +- Built structure is larger than person. +- Gear-like pressure is implied through shadow/geometry only. +- Traveler remains road-worn and secondary. +- Review file remains absent. + +## Prompt Handoff + +Carry Images 1-4 as visible-context references when possible. The final prompt must record `gpt-image-2` Advisor / host-native mode, exact source files read, and target `s28-v2-image.png`. + diff --git a/projects/2026-06-23-在路上/slides/s28/s28-v2-image.png b/projects/2026-06-23-在路上/slides/s28/s28-v2-image.png new file mode 100644 index 0000000..99d223f Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s28/s28-v2-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s28/s28-v2-prompt.md b/projects/2026-06-23-在路上/slides/s28/s28-v2-prompt.md new file mode 100644 index 0000000..4d02b7f --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s28/s28-v2-prompt.md @@ -0,0 +1,77 @@ +# S-28 Slide Prompt V2 + +## Status + +- Status: generation-ready. +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Overwrite allowed: no. + +## Based On Brief + +- Brief: `projects/2026-06-23-在路上/slides/s28/s28-v2-brief.md` +- Slide/version: `S-28 / s28-v2` +- Target image: `projects/2026-06-23-在路上/slides/s28/s28-v2-image.png` + +## Reference Inputs + +- Style module: B / high-angle city edge. +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; role: overall documentary road-film calibration, palette, restraint, avoid-drift. +- Image 2: `projects/2026-06-23-在路上/intake/imgs/B.高位观察型蓝图.png`; role: high-angle city-edge structure pressure, wet concrete, distant observational scale. +- Image 3: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png`; role: walking side silhouette continuity, coat, backpack, stride. +- Image 4: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`; role: walking-away silhouette continuity, backpack, coat length, non-heroic road-worn posture. +- Generation path note: actual reference images are attached in visible context as Image 1 through Image 4 for host-native generation. + +## Operation Notes + +- Task type: new image generation with references, not image editing. +- First read order: massive bridge / circular underpass shadow and wet city-edge ground before the traveler. +- Scale rule: traveler about 1/12 to 1/8 of frame, secondary to the built structure; no face detail. +- Structural metaphor rule: gear-like pressure must be implied only through real bridge, ramp, ring-shadow, drainage, and concrete geometry. + +## Advisor / Host-Native Record + +- `gpt-image-2` mode command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; recommendation `host-or-advisor`; `garden_mode_enabled: false`; `has_api_key: false`; model `gpt-image-2`. +- Interpreted mode for this task: Mode B host-native, because the host image-generation tool is available. +- Garden/API expected: no. +- Skill/template/reference files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.codex\skills\.system\imagegen\SKILL.md` + - `C:\Users\wangq\.codex\skills\.system\imagegen\references\prompting.md` +- Project/source files read: + - `AGENTS.md` + - `docs/workflows/slide-task-orchestration.md` + - `docs/workflows/slide-task-templates.md` + - `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s28-v2-slide-agent-packet.md` + - `projects/2026-06-23-在路上/slides/s28/s28-v2-brief.md` + - `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` + - `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + +## Final Prompt + +Use Image 1 as the overall documentary road-film style calibration. Use Image 2 as the high-angle city-edge / concrete pressure reference. Use Image 3 and Image 4 only to preserve the traveler's walking silhouette, long dark worn coat, backpack, stride, and road-worn body posture. Generate a new 16:9 horizontal photorealistic cinematic documentary still. + +Scene/backdrop: night or deep pre-dawn at the outside edge of a city, under or beside a massive concrete-and-steel bridge, circular ramp, underpass curve, or round structural shadow. Wet blue-gray ground, puddles, stains, rust traces, drain lines, retaining walls, hard urban texture. The built system must feel larger than the person. + +Subject: one small middle-aged male traveler, distant and secondary, about 1/12 to 1/8 of the frame. He is road-worn, wearing a long dark khaki-black coat and backpack consistent with Image 3 and Image 4, seen from a distant side/back walking angle. No face detail, no portrait emphasis, no hero pose. + +Composition/framing: high-angle or distant observational frame. The massive bridge / ring-shaped underpass geometry and its shadow dominate the image first. The traveler appears near the lower side or edge as a small moving coordinate crossing wet ground. The bridge shadow may suggest gear-like force through real circular geometry, repeated beams, drainage rings, and shadow teeth, but it must remain a believable urban structure, not a literal gear. + +Lighting/mood: cold low night or deep pre-dawn light, utilitarian and restrained. Coal-black shadows, cold gray concrete, wet dark blue-gray reflections, muted earth brown and dark khaki. No warm sunrise; no decorative neon. Mood is system pressure, continuation, and quiet endurance. + +Materials/textures: real wet concrete, pitted asphalt, puddle edges, rusty seams, mud stains, damp coat fabric, worn backpack, rough boots, industrial grime. Documentary realism with restrained poetic surrealism; the poetry comes from geometry and shadow, not fantasy effects. + +Output intent: production slide image for `S-28 / s28-v2`, road-film visual continuity with style module B. + +## Negative / Avoid + +No cyberpunk, neon, sci-fi UI, code, holograms, literal machine gears, fantasy machinery, clean CBD promotion, glass-tower advertisement, commercial skyline hero shot, travel advertising, fashion portrait, heroic centered stance, readable text, labels, logos, watermark, poster typography, extra people, vehicles as main subject, sunrise/sunset beauty shot, high saturation, or magical glow. + +## Output Intent + +- Save the generated host output as `projects/2026-06-23-在路上/slides/s28/s28-v2-image.png`. +- Do not write `s28-v2-review.md`. diff --git a/projects/2026-06-23-在路上/slides/s29/s29-v2-brief.md b/projects/2026-06-23-在路上/slides/s29/s29-v2-brief.md new file mode 100644 index 0000000..c741144 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s29/s29-v2-brief.md @@ -0,0 +1,34 @@ +# S-29 Brief V2 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Topology: direct-child, no Agent. +- Overwrite allowed: no. + +## Shot + +- Unit: `S-29`. +- Function: boot restarts one step, mud splashes lightly. +- Style module: character detail with A / muddy-road texture. +- First-read order: wet mud and boot motion before identity. + +## Visual Intent + +A low, close road-film detail of one worn boot taking the next step through shallow mud. The action should read as continuation after the heavier C-sequence pressure, not as a shoe advertisement or heroic stride. + +Keep the traveler present only through boot, lower trouser, coat hem edge, and motion. No face. No full-body portrait. Mud, water, grit, and small splash are the subject. + +## Reference Plan + +- Image 1: style calibration board; use for project palette and avoid-drift. +- Image 2: A landscape-led road reference; use for muddy-road continuity and cold terrain. +- Image 4: walking side character reference; use for worn boot, trouser, coat, stride, and road-worn identity cues. + +## Target Output + +- Prompt: `projects/2026-06-23-在路上/slides/s29/s29-v2-prompt.md` +- Image: `projects/2026-06-23-在路上/slides/s29/s29-v2-image.png` +- Review file: none. diff --git a/projects/2026-06-23-在路上/slides/s29/s29-v2-image.png b/projects/2026-06-23-在路上/slides/s29/s29-v2-image.png new file mode 100644 index 0000000..9ce8f38 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s29/s29-v2-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s29/s29-v2-prompt.md b/projects/2026-06-23-在路上/slides/s29/s29-v2-prompt.md new file mode 100644 index 0000000..042aeb6 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s29/s29-v2-prompt.md @@ -0,0 +1,79 @@ +# S-29 Slide Prompt V2 + +## Status + +- Status: ready for host-native generation. +- Slide/version: `S-29` / `s29-v2`. +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Topology: direct-child, no Agent. +- Overwrite allowed: no. + +## Based On Brief + +- Brief: `projects/2026-06-23-在路上/slides/s29/s29-v2-brief.md` +- Slide function: boot restarts one step, mud splashes lightly. +- First-read order: wet mud and boot motion before identity. + +## Reference Inputs + +- Style module: character detail with A / muddy-road texture. +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; role: overall documentary road-film calibration, low-saturation palette, anti-drift boundaries. +- Image 2: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`; role: landscape-led muddy-road continuity, cold terrain, wet ruts, frost and rough earth. +- Image 4: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png`; role: walking side character reference for worn boot, trouser, coat hem, stride, and road-worn identity cues. +- Character reference: Image 4 is required because the boot / lower-body detail belongs to the traveler. +- Generation path note: actual reference images are loaded in visible context for host-native generation; file paths are traceability, not edit targets. + +## Advisor / Host-Native Record + +- `gpt-image-2` mode command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `mode="B-or-C"`, `recommendation="host-or-advisor"`, `garden_mode_enabled=false`, `has_api_key=false`, `model="gpt-image-2"`. +- Interpreted mode for this task: Mode B host-native because the host image-generation tool is available. +- Garden/API expected: no. +- Skill/template/reference files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` + - `C:\Users\wangq\.codex\skills\.system\imagegen\SKILL.md` +- Project/source files read: + - `AGENTS.md` + - `VIDEO_WORKBENCH.md` + - `projects/2026-06-23-在路上/project.md` + - `projects/2026-06-23-在路上/execution-plan.md` + - `projects/2026-06-23-在路上/slides/slides.md` + - `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` + - `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + - `projects/2026-06-23-在路上/slides/s26/s26-v2-prompt.md` + - `projects/2026-06-23-在路上/slides/s27/s27-v2-prompt.md` + - `projects/2026-06-23-在路上/slides/s28/s28-v2-prompt.md` + +## Final Prompt + +Use case: photorealistic-natural. +Asset type: 16:9 cinematic documentary road-film still for slide S-29. +Input images: Image 1 is the overall style calibration board; Image 2 is the A muddy-road landscape reference; Image 4 is the walking side character reference for boot, trouser, coat hem, and stride. Use them as visible-context references, not edit targets. + +Create one 16:9 horizontal photorealistic cinematic documentary road-film still. The frame is a low close detail of the traveler restarting one step through shallow wet mud. First read must be real mud, water, grit, stones, road ruts, boot pressure, and a small restrained splash. Only after that should the viewer understand it is the same road-worn traveler continuing. + +Composition/framing: low ground-level camera, close to the road surface, focused on one worn boot entering or lifting from a muddy rut. The boot and lower trouser occupy about 20% to 35% of the frame; the rest is mud, shallow water, dirty road texture, stones, and the cold road surface. Include a small mud splash or water flick, but keep it physically modest and documentary. The coat hem may appear at the upper edge only as a dark road-worn shape. No face and no full body. + +Character handling: preserve the worn practical boot, dark trouser, long dark khaki-black coat edge, road-worn fabric, and walking motion cues from Image 4. The identity is conveyed through clothing and gait detail, not portrait likeness. + +Style/medium: photorealistic cinematic documentary still, restrained poetic realism. Use Image 1 for the project discipline: low saturation, road-film realism, no heroic framing. Use Image 2 for cold muddy-road continuity and physical terrain. The mood is continuation, not victory. + +Lighting/mood: cold pre-dawn or overcast dawn light, very low saturation. Palette: cold gray, wet blue-gray, earth brown, dark khaki, coal black, small pale highlights in water. No warm broad sunrise. + +Materials/textures: wet mud, dirty water, rough stones, boot leather or worn fabric, damp trouser cuff, tiny mud droplets, gritty road rut, dull reflected sky in puddles. + +Constraints: no text, no labels, no logos, no watermark. Keep it as an MV shot detail, not product photography. + +## Negative / Avoid + +No shoe advertisement, no clean product shot, no fashion photography, no heroic stride, no full-body traveler, no face, no portrait, no travel-ad road, no beautiful postcard sunrise, no high saturation, no fantasy splash, no slow-motion commercial water arc, no city neon, no UI, no readable text, no logos, no watermark. + +## Output Intent + +- Save the generated image as `projects/2026-06-23-在路上/slides/s29/s29-v2-image.png`. +- Keep `s29-v2-review.md` absent because review mode is none. diff --git a/projects/2026-06-23-在路上/slides/s30/s30-v2-brief.md b/projects/2026-06-23-在路上/slides/s30/s30-v2-brief.md new file mode 100644 index 0000000..35c3a9d --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s30/s30-v2-brief.md @@ -0,0 +1,34 @@ +# S-30 Brief V2 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Topology: direct-child, no Agent. +- Overwrite allowed: no. + +## Shot + +- Unit: `S-30`. +- Function: thin restrained dawn edge lights road. +- Style module: A / landscape-led road. +- First-read order: long muddy road and cold dawn edge before traveler. + +## Visual Intent + +A wide landscape-led frame where a very thin dawn edge starts to touch the muddy road. The shot must not become a postcard sunrise or motivational road image. + +The road remains larger than the person. If the traveler appears, he should be tiny, back-view or walking away, secondary to road and weather. + +## Reference Plan + +- Image 1: style calibration board; use for project palette and avoid-drift. +- Image 2: A landscape-led road reference; use for world-before-person composition and cold muddy road. +- Image 5: walking back character reference; use only if the small traveler is visible. + +## Target Output + +- Prompt: `projects/2026-06-23-在路上/slides/s30/s30-v2-prompt.md` +- Image: `projects/2026-06-23-在路上/slides/s30/s30-v2-image.png` +- Review file: none. diff --git a/projects/2026-06-23-在路上/slides/s30/s30-v2-image.png b/projects/2026-06-23-在路上/slides/s30/s30-v2-image.png new file mode 100644 index 0000000..1f8e6a0 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s30/s30-v2-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s30/s30-v2-prompt.md b/projects/2026-06-23-在路上/slides/s30/s30-v2-prompt.md new file mode 100644 index 0000000..9bb0604 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s30/s30-v2-prompt.md @@ -0,0 +1,77 @@ +# S-30 Slide Prompt V2 + +## Status + +- Status: ready for host-native generation. +- Slide/version: `S-30` / `s30-v2`. +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Topology: direct-child, no Agent. +- Overwrite allowed: no. + +## Based On Brief + +- Brief: `projects/2026-06-23-在路上/slides/s30/s30-v2-brief.md` +- Slide function: thin restrained dawn edge lights road. +- First-read order: long muddy road and cold dawn edge before traveler. + +## Reference Inputs + +- Style module: A / landscape-led road. +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; role: overall documentary road-film calibration, palette, restraint, avoid-drift. +- Image 2: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`; role: primary landscape-led muddy-road composition reference, world-before-person scale, cold dawn/pre-dawn terrain. +- Image 5: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`; role: small walking-away traveler silhouette, coat, backpack, posture. +- Character reference: Image 5 is used only for a tiny back-view traveler; do not turn this into a portrait. +- Generation path note: actual reference images are loaded in visible context for host-native generation; file paths are traceability, not edit targets. + +## Advisor / Host-Native Record + +- `gpt-image-2` mode command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `mode="B-or-C"`, `recommendation="host-or-advisor"`, `garden_mode_enabled=false`, `has_api_key=false`, `model="gpt-image-2"`. +- Interpreted mode for this task: Mode B host-native because the host image-generation tool is available. +- Garden/API expected: no. +- Skill/template/reference files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` + - `C:\Users\wangq\.codex\skills\.system\imagegen\SKILL.md` +- Project/source files read: + - `AGENTS.md` + - `VIDEO_WORKBENCH.md` + - `projects/2026-06-23-在路上/project.md` + - `projects/2026-06-23-在路上/execution-plan.md` + - `projects/2026-06-23-在路上/slides/slides.md` + - `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` + - `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + - `projects/2026-06-23-在路上/slides/s26/s26-v2-prompt.md` + - `projects/2026-06-23-在路上/slides/s27/s27-v2-prompt.md` + - `projects/2026-06-23-在路上/slides/s28/s28-v2-prompt.md` + +## Final Prompt + +Use case: photorealistic-natural. +Asset type: 16:9 cinematic documentary road-film still for slide S-30. +Input images: Image 1 is the overall style calibration board; Image 2 is the primary A landscape-led muddy-road reference; Image 5 is the optional tiny walking-away traveler reference for silhouette, coat, backpack, and posture. Use them as visible-context references, not edit targets. + +Create one 16:9 horizontal photorealistic cinematic documentary road-film still. The scene is a large cold muddy wilderness road at the thin edge of dawn. The first read must be the long road, wet ruts, frost or cold damp ground, rough earth, stones, puddles, and a narrow restrained dawn band touching the road. The person must not be the first read. + +Composition/framing: wide landscape-led road-film frame. The road begins near the lower frame and cuts toward distant open terrain. The horizon is low to mid, with cold sky and a very thin pale dawn line. If the traveler appears, make him very small, about 1/18 to 1/12 of the frame, back view or walking away on the muddy road, using Image 5 only for silhouette, coat, backpack, and posture. The traveler is a coordinate on the road, not a hero. + +Lighting/mood: restrained cold dawn, not a sunrise spectacle. A thin pale blue-gray or weak silver-yellow edge lights the road surface and puddles. Keep most of the image cold, muted, and damp. Warmth is barely present and should feel like endurance, not hope-poster triumph. + +Style/medium: photorealistic cinematic documentary still with restrained poetic realism. Use Image 1 for low-saturation project continuity and anti-drift discipline. Use Image 2 for landscape-led muddy-road structure, world-before-person scale, and cold terrain. The frame should feel like the MV is continuing through fatigue. + +Materials/textures: wet mud, shallow puddles, frost on grass or stones, rough road ruts, dark khaki earth, cold blue-gray sky, damp coat and backpack if visible. + +Constraints: no readable text, no subtitles, no logos, no UI, no watermark. Keep the shot quiet and physical. + +## Negative / Avoid + +No broad postcard sunrise, no sunset, no golden motivational road, no tourism landscape, no clean inspirational path, no hero backshot, no large portrait subject, no fashion photography, no high saturation, no dramatic god rays, no fantasy glow, no vehicles as the subject, no city neon, no readable text, no labels, no logos, no watermark. + +## Output Intent + +- Save the generated image as `projects/2026-06-23-在路上/slides/s30/s30-v2-image.png`. +- Keep `s30-v2-review.md` absent because review mode is none. diff --git a/projects/2026-06-23-在路上/slides/s31/s31-v2-brief.md b/projects/2026-06-23-在路上/slides/s31/s31-v2-brief.md new file mode 100644 index 0000000..8a61f4e --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s31/s31-v2-brief.md @@ -0,0 +1,34 @@ +# S-31 Brief V2 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Topology: direct-child, no Agent. +- Overwrite allowed: no. + +## Shot + +- Unit: `S-31`. +- Function: traveler reflection in puddle broken by water/wind. +- Style module: C / muddy starlight, with character trace. +- First-read order: dirty water, ripples, and broken reflection before traveler identity. + +## Visual Intent + +A low puddle/reflection shot where the traveler is present mainly as a broken, wind-disturbed reflection. It should stay physical: mud, water, ripple, grit, and cold reflected light. + +The traveler reflection should be legible but fragmented. No clean mirror lake, no full portrait, no fantasy portal. + +## Reference Plan + +- Image 1: style calibration board; use for project palette and avoid-drift. +- Image 3: repaired C muddy-starlight reference; use for real dirty water, broken reflections, and night/pre-dawn mud. +- Image 5: walking back character reference; use for reflected silhouette, coat, backpack, and posture. + +## Target Output + +- Prompt: `projects/2026-06-23-在路上/slides/s31/s31-v2-prompt.md` +- Image: `projects/2026-06-23-在路上/slides/s31/s31-v2-image.png` +- Review file: none. diff --git a/projects/2026-06-23-在路上/slides/s31/s31-v2-image.png b/projects/2026-06-23-在路上/slides/s31/s31-v2-image.png new file mode 100644 index 0000000..20cc6e1 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s31/s31-v2-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s31/s31-v2-prompt.md b/projects/2026-06-23-在路上/slides/s31/s31-v2-prompt.md new file mode 100644 index 0000000..b6353fb --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s31/s31-v2-prompt.md @@ -0,0 +1,77 @@ +# S-31 Slide Prompt V2 + +## Status + +- Status: ready for host-native generation. +- Slide/version: `S-31` / `s31-v2`. +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Topology: direct-child, no Agent. +- Overwrite allowed: no. + +## Based On Brief + +- Brief: `projects/2026-06-23-在路上/slides/s31/s31-v2-brief.md` +- Slide function: traveler reflection in puddle broken by water/wind. +- First-read order: dirty water, ripples, and broken reflection before traveler identity. + +## Reference Inputs + +- Style module: C / muddy starlight, with character trace. +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; role: overall documentary road-film calibration, low-saturation palette, anti-drift boundaries. +- Image 3: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png`; role: primary C muddy-water / broken-reflection reference; use for real wet mud, dirty puddle texture, cold reflected points, and physical ripple fragmentation. +- Image 5: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`; role: traveler silhouette reference for reflected back-view coat, backpack, and posture. +- Character reference: Image 5 is required because the traveler reflection must remain consistent, even though it is broken by water. +- Generation path note: actual reference images are loaded in visible context for host-native generation; file paths are traceability, not edit targets. + +## Advisor / Host-Native Record + +- `gpt-image-2` mode command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `mode="B-or-C"`, `recommendation="host-or-advisor"`, `garden_mode_enabled=false`, `has_api_key=false`, `model="gpt-image-2"`. +- Interpreted mode for this task: Mode B host-native because the host image-generation tool is available. +- Garden/API expected: no. +- Skill/template/reference files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` + - `C:\Users\wangq\.codex\skills\.system\imagegen\SKILL.md` +- Project/source files read: + - `AGENTS.md` + - `VIDEO_WORKBENCH.md` + - `projects/2026-06-23-在路上/project.md` + - `projects/2026-06-23-在路上/execution-plan.md` + - `projects/2026-06-23-在路上/slides/slides.md` + - `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` + - `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + - `projects/2026-06-23-在路上/slides/s26/s26-v2-prompt.md` + - `projects/2026-06-23-在路上/slides/s27/s27-v2-prompt.md` + - `projects/2026-06-23-在路上/slides/s28/s28-v2-prompt.md` + +## Final Prompt + +Use case: photorealistic-natural. +Asset type: 16:9 cinematic documentary road-film still for slide S-31. +Input images: Image 1 is the overall style calibration board; Image 3 is the primary C muddy-water / broken-reflection reference; Image 5 is the walking-back traveler reference for reflected coat, backpack, and posture. Use them as visible-context references, not edit targets. + +Create one 16:9 horizontal photorealistic cinematic documentary road-film still. The camera is very low beside a muddy puddle on the road at night or deep pre-dawn. The first read must be dirty water, wet mud, grit, stones, uneven puddle edge, wind ripples, and broken reflections. The viewer then discovers the traveler's reflection in the puddle, fragmented by moving water and mud. + +Composition/framing: ground-level puddle composition. The puddle and surrounding mud occupy most of the frame. The actual traveler may be outside the frame or only an indistinct edge above the puddle; the main human presence is the reflected back-view silhouette. The reflection should show a small broken dark coat shape and backpack shape consistent with Image 5, but warped and interrupted by ripples, stones, mud, and dirty water. Keep the reflection legible enough to read as the traveler, but never clean enough to become a mirror portrait. + +Lighting/mood: cold dark blue-gray pre-dawn or night. A few faint cold-white reflected points or pale sky fragments may appear in the water, but the scene must remain physical and muddy. Wind creates small ripple cuts through the reflection. Mood is continuation under disturbance, not revelation or triumph. + +Style/medium: photorealistic cinematic documentary road-film still with restrained poetic realism. Use Image 3 for muddy water, broken reflected light, wet grit, and low ground texture. Use Image 1 for low-saturation project continuity and anti-drift. Use Image 5 only to preserve the traveler's back-view coat, backpack, and posture through a broken reflection. + +Materials/textures: dirty water, wet mud, grit, stones, shallow puddle film, small ripples, cold light fragments, road rut edges, damp fabric only as reflection or border trace. + +Constraints: no readable text, no subtitles, no logos, no UI, no watermark. Keep it a road-film still, not fantasy art. + +## Negative / Avoid + +No clean mirror lake, no full-body traveler standing clearly above the puddle, no face, no portrait, no fashion reflection, no fantasy portal, no magic glow, no cosmic spectacle, no broad sunrise, no sunset, no warm horizon, no only-stars-without-mud, no only-mud-without-reflection, no shoe advertisement, no city neon, no UI, no readable text, no logos, no watermark. + +## Output Intent + +- Save the generated image as `projects/2026-06-23-在路上/slides/s31/s31-v2-image.png`. +- Keep `s31-v2-review.md` absent because review mode is none. diff --git a/projects/2026-06-23-在路上/slides/s32/s32-v2-brief.md b/projects/2026-06-23-在路上/slides/s32/s32-v2-brief.md new file mode 100644 index 0000000..b014c15 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s32/s32-v2-brief.md @@ -0,0 +1,71 @@ +# S-32 Slide Brief V2 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Version: `s32-v2`, resolved after scanning existing `s32-v1-image.png`. +- Write policy: append-only; overwrite allowed: no. +- Validation status after generation: generated / validation pending. + +## Source Basis + +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md`, row `S-32`. +- `projects/2026-06-23-在路上/slides/slides.md`, row `s32`. +- `projects/2026-06-23-在路上/visual-system/visual-system.md`. +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md`. +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md`. + +## Slide Function + +The line is “期待同行的你认领”. The frame should hint that another traveler or shared road exists without showing a clear companion. Use faint second footprints or a very weak far fire point as the signal. + +## First Read + +Mud, road texture, footprints, and distant darkness before any person. + +## Risk Class + +Environment / motif. No clear person should appear. + +## Reference Plan + +- Style module: D with motif handling, supported by A road-world restraint. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png`. +- Character reference: none; the source row explicitly says no clear person. +- Visible-context reference for generation: load the D style reference as `Image 1: restrained night-road fire / distant warmth style reference`. + +## Visual Requirements + +- 16:9 horizontal road-film still. +- Wet mud, dark road edge, broken footprints, and one ambiguous sign of another presence. +- The “other” must remain only a trace: blurry second footprints, a far pin of orange fire, or both. +- No companion body, no face, no group, no reunion. + +## Character Handling + +No visible character. Do not attach character identity references because they increase the chance of a person appearing where the shot requires absence. + +## Composition / Scale + +Low or medium-low observational frame. The road and muddy ground dominate. Distant fire, if present, should be tiny and far, about 1-2% of the frame. + +## Light / Color / Texture + +Dark wet blue-gray, earth brown, coal black, low saturation. Warm orange is allowed only as a faint point in the distance. Mud and footprints must feel real, not decorative. + +## Forbidden Drift + +No clear second person, no camp scene, no cozy fire, no group, no victory, no inspirational poster, no fantasy glow, no readable text, no UI, no logo, no watermark. + +## Acceptance Checklist + +- First read is mud / footprints / distant road, not a character. +- The hint of others remains ambiguous and restrained. +- Fire is a signal, not a destination. +- The image fits the accepted `在路上` documentary poetic-realism style. + +## Prompt Handoff + +Carry the D style module and global kernel into the prompt. The final prompt must state that the actual D style reference is loaded as `Image 1` for host-native generation. diff --git a/projects/2026-06-23-在路上/slides/s32/s32-v2-image.png b/projects/2026-06-23-在路上/slides/s32/s32-v2-image.png new file mode 100644 index 0000000..874676a Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s32/s32-v2-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s32/s32-v2-prompt.md b/projects/2026-06-23-在路上/slides/s32/s32-v2-prompt.md new file mode 100644 index 0000000..e9e354a --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s32/s32-v2-prompt.md @@ -0,0 +1,74 @@ +# S-32 Slide Prompt V2 + +## Status + +- Final prompt/spec prepared through `gpt-image-2` Skill Advisor / host-native workflow. +- Generation approval: task-level instruction, scope `S-32`, running mode `prompt and generate`. +- Validation status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s32/s32-v2-brief.md` + +## Reference Inputs + +- Style module: D, with A road-world restraint. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png`. +- Reference role: `Image 1: restrained night-road fire / distant warmth style reference`. +- Character references: none; this slide requires no clear person. +- Generation path note: use host-native `image_gen` after loading the real style reference into visible context. Ignore prior broad-run `s32-v1-image.png`. + +## Operation Notes + +- Append-only target: `projects/2026-06-23-在路上/slides/s32/s32-v2-image.png`. +- Do not create review, packet, or return files. +- Output should be 16:9 horizontal. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host provides `image_gen`. +- Garden/API expected: no. +- Skill files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Project sources read: + - `AGENTS.md` + - `VIDEO_WORKBENCH.md` + - `docs/workflows/project-lifecycle.md` + - `docs/workflows/slide-task-orchestration.md` + - `docs/workflows/slide-task-templates.md` + - `handoff/2026-06-25-on-the-road-formal-production-handoff.md` + - `projects/2026-06-23-在路上/project.md` + - `projects/2026-06-23-在路上/execution-plan.md` + - `projects/2026-06-23-在路上/slides/slides.md` + - `projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/task-plan.md` + - `projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/production-run-status.md` + - `projects/2026-06-23-在路上/visual-system/visual-system.md` + - `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` + - `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + - `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md` + +## Final Prompt + +Create a photorealistic cinematic documentary road-film still, 16:9 horizontal. + +Use Image 1 only as a restrained night-road fire / distant warmth style reference. Preserve the project mood: real documentary texture with restrained poetic realism, low saturation, environment before person, road before emotion. + +Scene: a dark wet muddy road at night or deep pre-dawn. The ground is the main subject: wet mud, rough stones, broken ruts, soft puddle edges, and two sets of footprints where one set is clearer and the other is faint, partial, and almost lost in the mud. Far in the distance there may be one extremely weak orange fire point, small enough to feel like a signal rather than a destination. + +Composition: low or medium-low observational frame. The road and mud dominate the frame. No clear human figure. The visual clue of “others on the road” should be ambiguous: blurry second footprints, a tiny far fire point, or both. Keep empty darkness and road depth around the trace. + +Lighting and color: wet dark blue-gray, coal black, earth brown, dark khaki; very low saturation. Warm orange appears only as a tiny far signal, not as a campfire scene. Real grit, mud thickness, moisture, and imperfect documentary texture. + +First-read order: mud and road texture first, footprints second, distant fire signal last. + +## Negative / Avoid + +No visible companion, no group, no face, no body silhouette, no cozy camp scene, no large fire, no reunion, no victory, no fantasy glow, no beautiful tourism landscape, no motivational poster, no sunrise, no sunset, no readable text, no labels, no logos, no watermark, no UI. + +## Output Intent + +Save the selected host-generated image as `projects/2026-06-23-在路上/slides/s32/s32-v2-image.png`. diff --git a/projects/2026-06-23-在路上/slides/s34/s34-v2-brief.md b/projects/2026-06-23-在路上/slides/s34/s34-v2-brief.md new file mode 100644 index 0000000..84f2ec2 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s34/s34-v2-brief.md @@ -0,0 +1,73 @@ +# S-34 Slide Brief V2 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Version: `s34-v2`, resolved after scanning existing `s34-v1-image.png`. +- Write policy: append-only; overwrite allowed: no. +- Validation status after generation: generated / validation pending. + +## Source Basis + +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md`, row `S-34`. +- `projects/2026-06-23-在路上/slides/slides.md`, row `s34`. +- `projects/2026-06-23-在路上/visual-system/visual-system.md`. +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md`. +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md`. + +## Slide Function + +The line is “踏破风化了的鞋跟”. Let the worn heel, mud, stones, and forward motion carry the traveler’s history without turning the image into a shoe advertisement. + +## First Read + +Old boot heel, mud, and rough road texture before the body. + +## Risk Class + +Object / body detail. + +## Reference Plan + +- Style module: A road texture with character-detail handling. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png`. +- Visible-context references for generation: + - `Image 1: walking side character slice, boot / coat / backpack continuity reference`. + - `Image 2: A muddy road / landscape-led style reference`. + +## Visual Requirements + +- Close low-angle detail of the old boot heel in wet mud and stones. +- The heel should be worn, cracked, and functional, not stylish. +- Show the next step continuing forward through mud. +- Body should be partial only: lower leg, coat hem, or backpack hint at most. + +## Character Handling + +Use the walking-side slice for boot, stride, coat, and backpack continuity. Do not make a portrait or full-body hero shot. + +## Composition / Scale + +Low ground-level close or near-close frame. Boot heel and mud occupy the main visual field; the road direction should still imply forward movement. + +## Light / Color / Texture + +Cold gray / earth brown / dark khaki, wet mud, grit, scratched leather, worn rubber, small splashes, real stones. No clean studio light. + +## Forbidden Drift + +No product ad, no new clean shoe, no brand logo, no fashion boot, no heroic full body, no clean scenic road, no fantasy glow, no text. + +## Acceptance Checklist + +- The image reads as cost and continuation, not product display. +- The old boot belongs to the established traveler. +- Mud and stones feel documentary and tactile. +- The action continues forward. + +## Prompt Handoff + +Carry the A style module, global style kernel, and walking-side character reference into the prompt. The final prompt must state that the actual references are loaded as `Image 1` and `Image 2`. diff --git a/projects/2026-06-23-在路上/slides/s34/s34-v2-image.png b/projects/2026-06-23-在路上/slides/s34/s34-v2-image.png new file mode 100644 index 0000000..c12f7ed Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s34/s34-v2-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s34/s34-v2-prompt.md b/projects/2026-06-23-在路上/slides/s34/s34-v2-prompt.md new file mode 100644 index 0000000..35c281c --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s34/s34-v2-prompt.md @@ -0,0 +1,61 @@ +# S-34 Slide Prompt V2 + +## Status + +- Final prompt/spec prepared through `gpt-image-2` Skill Advisor / host-native workflow. +- Generation approval: task-level instruction, scope `S-34`, running mode `prompt and generate`. +- Validation status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s34/s34-v2-brief.md` + +## Reference Inputs + +- Style module: A plus character-detail handling. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png`. +- Reference roles: + - `Image 1: walking side character slice, boot / stride / coat / backpack continuity reference`. + - `Image 2: A muddy road / landscape-led style reference`. +- Generation path note: use host-native `image_gen` after loading the real image references into visible context. Ignore prior broad-run `s34-v1-image.png`. + +## Operation Notes + +- Append-only target: `projects/2026-06-23-在路上/slides/s34/s34-v2-image.png`. +- Do not create review, packet, or return files. +- Output should be 16:9 horizontal. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host provides `image_gen`. +- Garden/API expected: no. +- Skill files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Project sources read: same production-light source set listed in `s32-v2-prompt.md`, plus the `S-34` source row and assigned slide folder. + +## Final Prompt + +Create a photorealistic cinematic documentary road-film still, 16:9 horizontal. + +Use Image 1 as the character continuity reference for the traveler’s walking-side boot, stride, coat hem, and road-worn outfit. Use Image 2 as the muddy-road style and palette reference. Do not copy either image literally; adapt them into a new low-angle boot-detail shot. + +Scene: a close low ground-level view of an old boot heel stepping forward through wet mud and rough stones on a long road. The heel is worn down, cracked, dirty, and weathered by distance. Mud clings to the sole and splashes lightly as the next step continues. A small part of the dark coat hem or lower leg may enter the frame, but the boot, mud, stones, and forward motion are the subject. + +Composition: boot heel and muddy ground dominate the frame. The road direction should lead forward into soft distance without becoming a scenic travel image. Keep the human body partial and secondary. The image should feel like history carried by an object. + +Lighting and color: cold gray, earth brown, dark khaki, wet dark blue-gray, low saturation. Overcast or pre-dawn diffuse light. Real leather wear, rubber abrasion, grit, stones, wet mud thickness, imperfect documentary texture. + +First-read order: worn heel first, mud and stones second, continuing step third. + +## Negative / Avoid + +No shoe advertisement, no clean new boot, no visible brand logo, no fashion photography, no polished product lighting, no full-body hero pose, no face, no inspirational poster, no bright sunrise, no fantasy glow, no readable text, no watermark, no UI. + +## Output Intent + +Save the selected host-generated image as `projects/2026-06-23-在路上/slides/s34/s34-v2-image.png`. diff --git a/projects/2026-06-23-在路上/slides/s35/s35-v1-brief.md b/projects/2026-06-23-在路上/slides/s35/s35-v1-brief.md new file mode 100644 index 0000000..a874a46 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s35/s35-v1-brief.md @@ -0,0 +1,73 @@ +# S-35 Slide Brief V1 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Version: `s35-v1`, resolved after scanning an empty `s35` folder. +- Write policy: append-only; overwrite allowed: no. +- Validation status after generation: generated / validation pending. + +## Source Basis + +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md`, row `S-35`. +- `projects/2026-06-23-在路上/slides/slides.md`, row `s35`. +- `projects/2026-06-23-在路上/visual-system/visual-system.md`. +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md`. +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md`. + +## Slide Function + +The line is “我不再迷惘”. Direction should form through action: the traveler pauses and turns toward a darker but clearer road. Do not make it a revelation or motivational victory. + +## First Read + +Forked road and chosen direction before the person. + +## Risk Class + +Medium human / back view. + +## Reference Plan + +- Style module: A landscape-led road. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`. +- Visible-context references for generation: + - `Image 1: walking back character slice, silhouette / coat / backpack continuity reference`. + - `Image 2: A muddy road / landscape-led style reference`. + +## Visual Requirements + +- Medium-wide or wide shot. +- A fork or bend in the rough road: one route may be lighter/uncertain, one darker but clearer. +- The traveler is seen from back or three-quarter back, small to medium-small, turning into action rather than posing. +- No glow of sudden enlightenment. + +## Character Handling + +Use walking-back reference for silhouette, coat, backpack, and road-worn posture. Face should not be visible. + +## Composition / Scale + +Environment first. Traveler about 1/8 to 1/6 of the frame, placed off-center. The chosen road line should carry the eye forward. + +## Light / Color / Texture + +Cold dawn or deep pre-dawn haze; low saturation. Mud, frost, stones, wind, rough road texture. A faint directional light may clarify the path but must stay restrained. + +## Forbidden Drift + +No hero stance, no arms-open posture, no bright revelation, no sunrise explosion, no clean road, no travel ad, no portrait MV, no text. + +## Acceptance Checklist + +- Direction forms through movement, not symbolism alone. +- The darker road is clear enough to read as chosen. +- Traveler remains a coordinate inside the road-world. +- Mood is steadier, not victorious. + +## Prompt Handoff + +Carry A module and walking-back character reference into the prompt. The final prompt must state that actual references are loaded as `Image 1` and `Image 2`. diff --git a/projects/2026-06-23-在路上/slides/s35/s35-v1-image.png b/projects/2026-06-23-在路上/slides/s35/s35-v1-image.png new file mode 100644 index 0000000..a239721 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s35/s35-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s35/s35-v1-prompt.md b/projects/2026-06-23-在路上/slides/s35/s35-v1-prompt.md new file mode 100644 index 0000000..9e91127 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s35/s35-v1-prompt.md @@ -0,0 +1,61 @@ +# S-35 Slide Prompt V1 + +## Status + +- Final prompt/spec prepared through `gpt-image-2` Skill Advisor / host-native workflow. +- Generation approval: task-level instruction, scope `S-35`, running mode `prompt and generate`. +- Validation status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s35/s35-v1-brief.md` + +## Reference Inputs + +- Style module: A. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`. +- Reference roles: + - `Image 1: walking back character slice, silhouette / coat / backpack continuity reference`. + - `Image 2: A muddy road / landscape-led style reference`. +- Generation path note: use host-native `image_gen` after loading the real image references into visible context. + +## Operation Notes + +- Append-only target: `projects/2026-06-23-在路上/slides/s35/s35-v1-image.png`. +- Do not create review, packet, or return files. +- Output should be 16:9 horizontal. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host provides `image_gen`. +- Garden/API expected: no. +- Skill files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Project sources read: same production-light source set listed in `s32-v2-prompt.md`, plus the `S-35` source row and assigned slide folder. + +## Final Prompt + +Create a photorealistic cinematic documentary road-film still, 16:9 horizontal. + +Use Image 1 as the traveler silhouette continuity reference: back view, long dark coat, old backpack, road-worn posture, still capable of walking. Use Image 2 as the muddy wilderness road style reference. Create a new scene, not a literal copy. + +Scene: a rough muddy road forks or bends in a cold open landscape. The traveler has paused for a moment and is turning toward a darker but clearer route. The chosen road should feel difficult but legible, not magically bright. Mud, stones, ruts, frost, and wind carry the reality of the place. + +Composition: medium-wide to wide frame. Road and fork dominate before the person. The traveler is back view or three-quarter back, off-center, about 1/8 to 1/6 of the frame, caught in the action of turning and beginning to move. The eye follows the chosen road forward into cold distance. + +Lighting and color: overcast dawn or deep pre-dawn haze, low saturation, cold gray, earth brown, dark khaki, wet dark blue-gray. A thin restrained directional light may separate the road edge, but no dramatic revelation. + +First-read order: forked road first, chosen direction second, traveler movement third. + +## Negative / Avoid + +No heroic pose, no arms-open stance, no bright enlightenment beam, no commercial sunrise, no postcard landscape, no clean road, no portrait MV, no face close-up, no motivational poster, no readable text, no logos, no watermark, no UI. + +## Output Intent + +Save the selected host-generated image as `projects/2026-06-23-在路上/slides/s35/s35-v1-image.png`. diff --git a/projects/2026-06-23-在路上/slides/s36/s36-v1-brief.md b/projects/2026-06-23-在路上/slides/s36/s36-v1-brief.md new file mode 100644 index 0000000..59cd68f --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s36/s36-v1-brief.md @@ -0,0 +1,73 @@ +# S-36 Slide Brief V1 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Version: `s36-v1`, resolved after scanning an empty `s36` folder. +- Write policy: append-only; overwrite allowed: no. +- Validation status after generation: generated / validation pending. + +## Source Basis + +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md`, row `S-36`. +- `projects/2026-06-23-在路上/slides/slides.md`, row `s36`. +- `projects/2026-06-23-在路上/visual-system/visual-system.md`. +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md`. +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md`. + +## Slide Function + +The line is “哪怕前方依然陌生”. The route remains unresolved: the road goes into darkness, wind, and sand, and the traveler keeps walking into the unknown. + +## First Read + +Dark road entering wind / sand / unknown space before the traveler. + +## Risk Class + +Distant human / back view. + +## Reference Plan + +- Style module: A landscape-led road. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`. +- Visible-context references for generation: + - `Image 1: walking back character slice, small back-view continuity reference`. + - `Image 2: A muddy road / landscape-led style reference`. + +## Visual Requirements + +- Wide or distant shot. +- Road moves into dark wind, dust, or sand; the end remains unknown. +- Traveler continues forward, back view, small and secondary. +- Preserve uncertainty; do not resolve the image with warmth or victory. + +## Character Handling + +Use walking-back reference for silhouette and backpack/coat continuity. Face should not be visible. + +## Composition / Scale + +Environment dominates. Traveler about 1/10 to 1/7 of the frame, smaller if the road and storm need more scale. Road should pull into darkness. + +## Light / Color / Texture + +Cold gray, dark khaki, earth brown, wet blue-gray, wind-blown grit. Low contrast but readable road line. No bright sky. + +## Forbidden Drift + +No clean desert fantasy, no heroic lone-wanderer poster, no sunrise/sunset resolution, no action-movie storm, no sci-fi, no UI, no text. + +## Acceptance Checklist + +- Unknown remains visible and unresolved. +- The traveler is still moving, not posing. +- Landscape pressure is larger than the person. +- Mood is steady and difficult, not triumphant. + +## Prompt Handoff + +Carry A module and walking-back character reference into the prompt. The final prompt must state that actual references are loaded as `Image 1` and `Image 2`. diff --git a/projects/2026-06-23-在路上/slides/s36/s36-v1-image.png b/projects/2026-06-23-在路上/slides/s36/s36-v1-image.png new file mode 100644 index 0000000..989827e Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s36/s36-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s36/s36-v1-prompt.md b/projects/2026-06-23-在路上/slides/s36/s36-v1-prompt.md new file mode 100644 index 0000000..313a100 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s36/s36-v1-prompt.md @@ -0,0 +1,61 @@ +# S-36 Slide Prompt V1 + +## Status + +- Final prompt/spec prepared through `gpt-image-2` Skill Advisor / host-native workflow. +- Generation approval: task-level instruction, scope `S-36`, running mode `prompt and generate`. +- Validation status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s36/s36-v1-brief.md` + +## Reference Inputs + +- Style module: A. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`. +- Reference roles: + - `Image 1: walking back character slice, small back-view continuity reference`. + - `Image 2: A muddy road / landscape-led style reference`. +- Generation path note: use host-native `image_gen` after loading the real image references into visible context. + +## Operation Notes + +- Append-only target: `projects/2026-06-23-在路上/slides/s36/s36-v1-image.png`. +- Do not create review, packet, or return files. +- Output should be 16:9 horizontal. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host provides `image_gen`. +- Garden/API expected: no. +- Skill files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Project sources read: same production-light source set listed in `s32-v2-prompt.md`, plus the `S-36` source row and assigned slide folder. + +## Final Prompt + +Create a photorealistic cinematic documentary road-film still, 16:9 horizontal. + +Use Image 1 as the small back-view traveler continuity reference: dark long coat, old backpack, road-worn walking silhouette. Use Image 2 as the muddy wilderness road style reference. Create a new wide shot, not a literal copy. + +Scene: a rough road enters a dark unknown stretch of wind, dust, and cold haze. The foreground has wet mud, stones, ruts, and gritty road texture; farther ahead the road is swallowed by darker wind and sand. The traveler keeps walking into it, back view, small and secondary. + +Composition: wide or distant frame. The road and unknown distance dominate the image. The traveler should be about 1/10 to 1/7 of the frame, placed along the road line, still moving forward. Keep the final destination invisible. + +Lighting and color: cold gray, dark khaki, coal black, earth brown, wet blue-gray. Low saturation, overcast or deep pre-dawn, wind-blown grit in the air. The road line remains readable but the future remains unresolved. + +First-read order: road into darkness first, wind and unknown second, tiny walking traveler third. + +## Negative / Avoid + +No resolved destination, no bright sunrise, no sunset glow, no heroic centered silhouette, no action-movie dust storm, no fantasy wasteland, no sci-fi, no clean travel ad, no portrait MV, no readable text, no logos, no watermark, no UI. + +## Output Intent + +Save the selected host-generated image as `projects/2026-06-23-在路上/slides/s36/s36-v1-image.png`. diff --git a/projects/2026-06-23-在路上/slides/s37/s37-v1-brief.md b/projects/2026-06-23-在路上/slides/s37/s37-v1-brief.md new file mode 100644 index 0000000..3498b45 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s37/s37-v1-brief.md @@ -0,0 +1,76 @@ +# S-37 Slide Brief V1 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Version: `s37-v1`, resolved after scanning an empty `s37` folder. +- Write policy: append-only; overwrite allowed: no. +- Validation status after generation: generated / validation pending. + +## Source Basis + +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md`, row `S-37`. +- `projects/2026-06-23-在路上/slides/slides.md`, row `s37`. +- `projects/2026-06-23-在路上/visual-system/visual-system.md`. +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md`. +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md`. + +## Slide Function + +The line is “在这条路上”. The road becomes the protagonist again: a long road crosses the frame, the person is at the edge, and faint fire far away suggests continuation. + +## First Read + +Long road / landscape before person and fire. + +## Risk Class + +Distant human / landscape-led frame. + +## Reference Plan + +- Style module: A / D. +- Active style references: + - `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png` + - `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`. +- Visible-context references for generation: + - `Image 1: walking back character slice, tiny edge traveler silhouette reference`. + - `Image 2: A landscape-led road style reference`. + - `Image 3: D distant fire / restrained warmth style reference`. + +## Visual Requirements + +- Very wide or wide landscape-led road frame. +- Long road crosses or cuts through the frame; road is the subject. +- Traveler at the edge, tiny or small, walking forward. +- Faint far fire point or points are allowed, but they must not become a camp scene. + +## Character Handling + +Use walking-back reference only to stabilize silhouette, coat, and backpack. The character should be small enough that identity reads mainly through outline. + +## Composition / Scale + +Road dominates. Traveler about 1/12 to 1/8 of the frame, near an edge or lower-side area. Fire point should be far and tiny. + +## Light / Color / Texture + +Cold dark road, wet blue-gray, earth brown, dark khaki, low saturation. Small warm orange far away only as a signal. + +## Forbidden Drift + +No centered hero, no large campfire, no group, no arrival, no victory, no postcard horizon, no bright sunrise, no text. + +## Acceptance Checklist + +- The road is the protagonist. +- Person is pushed back to edge / scale. +- Fire remains a distant signal. +- The shot bridges the action into later D fire material without resolving the story. + +## Prompt Handoff + +Carry A/D modules, walking-back character reference, and global kernel into the prompt. The final prompt must state that actual references are loaded as `Image 1`, `Image 2`, and `Image 3`. diff --git a/projects/2026-06-23-在路上/slides/s37/s37-v1-image.png b/projects/2026-06-23-在路上/slides/s37/s37-v1-image.png new file mode 100644 index 0000000..fec947f Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s37/s37-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s37/s37-v1-prompt.md b/projects/2026-06-23-在路上/slides/s37/s37-v1-prompt.md new file mode 100644 index 0000000..3a6a2ec --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s37/s37-v1-prompt.md @@ -0,0 +1,64 @@ +# S-37 Slide Prompt V1 + +## Status + +- Final prompt/spec prepared through `gpt-image-2` Skill Advisor / host-native workflow. +- Generation approval: task-level instruction, scope `S-37`, running mode `prompt and generate`. +- Validation status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s37/s37-v1-brief.md` + +## Reference Inputs + +- Style module: A / D. +- Active style references: + - `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png` + - `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`. +- Reference roles: + - `Image 1: walking back character slice, tiny edge traveler silhouette reference`. + - `Image 2: A landscape-led road style reference`. + - `Image 3: D distant fire / restrained warmth style reference`. +- Generation path note: use host-native `image_gen` after loading the real image references into visible context. + +## Operation Notes + +- Append-only target: `projects/2026-06-23-在路上/slides/s37/s37-v1-image.png`. +- Do not create review, packet, or return files. +- Output should be 16:9 horizontal. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host provides `image_gen`. +- Garden/API expected: no. +- Skill files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Project sources read: same production-light source set listed in `s32-v2-prompt.md`, plus the `S-37` source row and assigned slide folder. + +## Final Prompt + +Create a photorealistic cinematic documentary road-film still, 16:9 horizontal. + +Use Image 1 as the tiny back-view traveler silhouette reference: dark long coat, old backpack, road-worn walking outline. Use Image 2 as the landscape-led road composition reference. Use Image 3 only for restrained distant fire warmth. Create a new wide shot, not a literal copy. + +Scene: a long wet road crosses or cuts through a large cold landscape at night or deep pre-dawn. The road is the protagonist: mud, ruts, stones, wet dark surface, wind, and distance. The traveler is at the edge of the frame, tiny or small, still walking. Far ahead, one or a few weak orange fire points appear as distant signals, not as a camp or destination. + +Composition: very wide or wide landscape-led frame. Environment first, road second, faint fire third, traveler last. Put the traveler near an edge or lower-side area, about 1/12 to 1/8 of the frame. The road line must carry the image across the frame and into distance. + +Lighting and color: low saturation, cold gray, wet blue-gray, earth brown, dark khaki, coal black. Warm orange appears only as tiny far fire points. Real documentary texture, restrained poetic realism. + +First-read order: road / landscape first, distant fire signal second, tiny traveler at edge third. + +## Negative / Avoid + +No centered hero, no large person, no portrait, no large campfire, no tent, no camping chairs, no group, no reunion, no arrival, no victory, no postcard sunrise, no bright sunset, no motivational poster, no readable text, no logos, no watermark, no UI. + +## Output Intent + +Save the selected host-generated image as `projects/2026-06-23-在路上/slides/s37/s37-v1-image.png`. diff --git a/projects/2026-06-23-在路上/slides/s38/s38-v1-brief.md b/projects/2026-06-23-在路上/slides/s38/s38-v1-brief.md new file mode 100644 index 0000000..cbf6ba1 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s38/s38-v1-brief.md @@ -0,0 +1,72 @@ +# S-38 Slide Brief V1 + +## Status + +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Version resolved by folder scan: `s38-v1` +- Output status target: generated / validation pending + +## Source Basis + +- `intake/阶段五-镜头清单骨架.md`: S-38, "燃烧寂静的篝火啊"; small roadside fire at night with the traveler alone. +- `slides/slides.md`: small night-road fire with traveler alone; style anchor D; priority. +- `visual-system/visual-system.md` +- `visual-system/style/style-prompt-master.md` +- `visual-system/characters/character-reference-registry.md` + +## Slide Function + +Emotion warms for the first time, but the image must stay quiet and solitary, not lively or communal. + +## First Read + +Night road and a restrained small fire first; traveler second; backpack, mud, and road texture third. + +## Risk Class + +medium human, fire motif + +## Reference Plan + +- Style module: D Night Road Fire. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png` +- Visible-context roles for generation: Image 1 style reference, Image 2 walking/back identity reference. + +## Visual Requirements + +- 16:9 horizontal cinematic documentary still. +- Small roadside fire, about 3%-6% of frame. +- Traveler alone, small to medium-small, back or three-quarter back, not facing camera. +- Firelight touches backpack, coat edge, boots, and wet mud. +- Road continues beyond the fire. + +## Character Handling + +Use the active walking/back reference for coat length, backpack shape, boots, and posture. Identity is silhouette-led, not face-led. + +## Composition / Scale + +Medium-wide or wide road-film frame. Keep the environment larger than the person; traveler about 1/8 to 1/6 of frame. + +## Light / Color / Texture + +Dark wet blue-gray night, earth-brown mud, coal-black coat, restrained local warm-orange firelight. + +## Forbidden Drift + +No camping setup, tents, chairs, cookware, group gathering, victory pose, cozy campsite, large theatrical bonfire, readable text, logo, or watermark. + +## Acceptance Checklist + +- Fire is a quiet emotional anchor, not a camp scene. +- Traveler remains road-worn and capable. +- Road continues beyond the fire. +- Image status remains validation pending for parent/user confirmation. + +## Prompt Handoff + +Build the final prompt through the `gpt-image-2` Advisor / host-native workflow using the D module and the walking/back character reference. diff --git a/projects/2026-06-23-在路上/slides/s38/s38-v1-image.png b/projects/2026-06-23-在路上/slides/s38/s38-v1-image.png new file mode 100644 index 0000000..7049383 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s38/s38-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s38/s38-v1-prompt.md b/projects/2026-06-23-在路上/slides/s38/s38-v1-prompt.md new file mode 100644 index 0000000..614a811 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s38/s38-v1-prompt.md @@ -0,0 +1,73 @@ +# S-38 Slide Prompt V1 + +## Status + +- Ready for host-native image generation. +- Review mode: none. +- Target slide status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s38/s38-v1-brief.md` + +## Reference Inputs + +- Style module: D / Night Road Fire. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png` +- Generation visible-context roles: Image 1 = D style reference; Image 2 = walking/back identity reference. + +## Operation Notes + +- Project-bound prompt produced through `gpt-image-2` Skill Advisor / host-native workflow. +- Garden/API expected: no. +- Actual image references must be loaded into visible context before host generation. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host has built-in `image_gen`. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Source files read: + - `AGENTS.md` + - `VIDEO_WORKBENCH.md` + - `docs/workflows/project-lifecycle.md` + - `docs/workflows/slide-task-orchestration.md` + - `docs/workflows/slide-task-templates.md` + - `handoff/2026-06-25-on-the-road-formal-production-handoff.md` + - `projects/2026-06-23-在路上/project.md` + - `projects/2026-06-23-在路上/execution-plan.md` + - `projects/2026-06-23-在路上/slides/slides.md` + - `projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/task-plan.md` + - `projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/production-run-status.md` + - `projects/2026-06-23-在路上/visual-system/visual-system.md` + - `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` + - `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + - relevant S-38 row from `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md` + +## Final Prompt + +Use case: photorealistic-natural +Asset type: 16:9 MV production still for slide S-38. +Input images: Image 1 is the D night-road fire style reference; Image 2 is the walking/back identity reference for the traveler. +Primary request: Create a photorealistic cinematic documentary road-film still of a small roadside fire burning quietly on a dark wet night road, with one road-worn male traveler alone near it. +Scene/backdrop: dark wet wilderness road at night or deep pre-dawn, mud, stones, damp ruts, low roadside darkness, road continuing beyond the fire. +Subject: a 50-year-old male traveler, seen from back or three-quarter back, dark long coat, old backpack, old boots, steady but tired posture. Use Image 2 for coat, backpack, boot, and walking/back silhouette continuity. Do not make him a portrait subject. +Style reference: use Image 1 for restrained local firelight, dark road mood, and anti-camping boundary. +Composition/framing: medium-wide 16:9 frame; environment first, road second, fire third, traveler fourth. Traveler about 1/8 to 1/6 of frame. Fire about 3% to 6% of frame, near the roadside and not centered like a camp scene. +Lighting/mood: cold wet blue-gray night with a small warm-orange fire signal. Firelight softly touches backpack, coat edge, boots, and muddy ground. Quiet, solitary, unfinished warmth; no triumph. +Color palette: low saturation cold gray, earth brown, coal black, wet dark blue-gray, restrained orange fire. +Textures: real mud, wet stones, coat fabric wear, old backpack, worn boot edges, light smoke, damp road grit. +Constraints: world larger than person; road longer than person; traveler is only the coordinate that keeps moving. No text. + +## Negative / Avoid + +Avoid camping advertisement, tents, camping chairs, cookware, picnic setup, cozy campsite, group gathering, reunion, victory pose, theatrical bonfire, heroic centered stance, fashion travel photography, bright sunset, high saturation, readable text, labels, logos, watermarks, UI. + +## Output Intent + +Save selected host output as `projects/2026-06-23-在路上/slides/s38/s38-v1-image.png` and record status as generated / validation pending. diff --git a/projects/2026-06-23-在路上/slides/s39/s39-v1-brief.md b/projects/2026-06-23-在路上/slides/s39/s39-v1-brief.md new file mode 100644 index 0000000..b854b60 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s39/s39-v1-brief.md @@ -0,0 +1,71 @@ +# S-39 Slide Brief V1 + +## Status + +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Version resolved by folder scan: `s39-v1` +- Output status target: generated / validation pending + +## Source Basis + +- `intake/阶段五-镜头清单骨架.md`: S-39, guitar solo beginning; no person, wind over muddy road and puddle. +- `slides/slides.md`: breath and silence; style A / C. +- `visual-system/visual-system.md` +- `visual-system/style/style-prompt-master.md` + +## Slide Function + +Release the chorus emotion into silence. The frame should breathe without adding narrative action. + +## First Read + +Mud, road, puddle, wind movement; no traveler. + +## Risk Class + +environment-only, puddle motif + +## Reference Plan + +- Style modules: A Landscape-Led Road + C Muddy Starlight restraint. +- Active style references: + - `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png` + - `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png` +- Character reference: none; the shot explicitly has no person. +- Visible-context roles for generation: Image 1 A style reference, Image 2 C repaired puddle reference. + +## Visual Requirements + +- No person, vehicle, animal, tent, sign, or readable text. +- Wind should be felt through small ripples, wet grass, dust, or rain trace, not illustrated as literal graphic lines. +- Muddy road and puddle stay documentary and restrained. + +## Character Handling + +None. Do not attach character references and do not insert a traveler. + +## Composition / Scale + +Static observational 16:9 frame, low or waist-high view along muddy road. Use quiet negative space and minimal motion cues. + +## Light / Color / Texture + +Cold gray, earth brown, wet blue-gray; faint pre-dawn or night ambient light. Puddle should have subtle darkness and texture. + +## Forbidden Drift + +No scenic tourism, no inspirational road, no clean mirror lake, no magic glow, no dramatic storm spectacle. + +## Acceptance Checklist + +- First read is breath/silence, not event. +- No person appears. +- Puddle and mud remain real. +- Image status remains validation pending for parent/user confirmation. + +## Prompt Handoff + +Build the final prompt through the `gpt-image-2` Advisor / host-native workflow using A and C style references only. diff --git a/projects/2026-06-23-在路上/slides/s39/s39-v1-image.png b/projects/2026-06-23-在路上/slides/s39/s39-v1-image.png new file mode 100644 index 0000000..e21db4f Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s39/s39-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s39/s39-v1-prompt.md b/projects/2026-06-23-在路上/slides/s39/s39-v1-prompt.md new file mode 100644 index 0000000..a3a96ca --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s39/s39-v1-prompt.md @@ -0,0 +1,59 @@ +# S-39 Slide Prompt V1 + +## Status + +- Ready for host-native image generation. +- Review mode: none. +- Target slide status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s39/s39-v1-brief.md` + +## Reference Inputs + +- Style modules: A / Landscape-Led Road; C / Muddy Starlight. +- Active style references: + - `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png` + - `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png` +- Character reference: none; no person should appear. +- Generation visible-context roles: Image 1 = A landscape-road style reference; Image 2 = C repaired puddle texture reference. + +## Operation Notes + +- Project-bound prompt produced through `gpt-image-2` Skill Advisor / host-native workflow. +- Garden/API expected: no. +- Actual image references must be loaded into visible context before host generation. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host has built-in `image_gen`. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Source files read: same project/task/visual-system files listed in `s38-v1-prompt.md`, plus the relevant S-39 row from `intake/阶段五-镜头清单骨架.md`. + +## Final Prompt + +Use case: photorealistic-natural +Asset type: 16:9 MV production still for slide S-39. +Input images: Image 1 is the A landscape-led muddy road reference; Image 2 is the C repaired muddy starlight / puddle texture reference. +Primary request: Create a quiet photorealistic cinematic documentary still of wind moving over a muddy road and a shallow puddle at the start of a guitar solo. No person appears. +Scene/backdrop: cold muddy road, wet ruts, stones, sparse roadside grass, shallow puddle with subtle ripples, damp grit and low empty space. Use Image 1 for landscape-led road composition and Image 2 for real mud and water texture. +Subject: the road and puddle are the subject. The frame should feel like breath and silence after an emotional release. +Composition/framing: static observational 16:9 frame, low or waist-high view along the muddy road. Keep the world open and sparse; no central figure. +Lighting/mood: cold gray pre-dawn or night-edge ambient light, restrained and documentary, with only tiny surface highlights in the puddle. +Color palette: low saturation cold gray, earth brown, wet blue-gray, coal-black shadows. +Textures: real mud, water ripples, grit, wet stones, sparse grass bending in wind, imperfect road surface. +Constraints: no person, no vehicle, no animal, no sign, no readable text. The first read must be silence, road, puddle, and wind. + +## Negative / Avoid + +Avoid travel advertising, postcard sunrise, heroic road, clean mirror lake, fantasy magic, glowing portal, dramatic storm spectacle, high saturation, readable text, labels, logos, watermarks, UI, any traveler silhouette. + +## Output Intent + +Save selected host output as `projects/2026-06-23-在路上/slides/s39/s39-v1-image.png` and record status as generated / validation pending. diff --git a/projects/2026-06-23-在路上/slides/s40/s40-v1-brief.md b/projects/2026-06-23-在路上/slides/s40/s40-v1-brief.md new file mode 100644 index 0000000..60e7b66 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s40/s40-v1-brief.md @@ -0,0 +1,73 @@ +# S-40 Slide Brief V1 + +## Status + +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Version resolved by folder scan: `s40-v1` +- Output status target: generated / validation pending + +## Source Basis + +- `intake/阶段五-镜头清单骨架.md`: S-40; high wide shot, tiny person slowly moving along a deserted road. +- `slides/slides.md`: documentary observation returns; style A / B. +- `visual-system/visual-system.md` +- `visual-system/style/style-prompt-master.md` +- `visual-system/characters/character-reference-registry.md` + +## Slide Function + +Reopen the inner space by making the world large and the traveler small again. + +## First Read + +Large terrain and road pattern first; tiny moving traveler second. + +## Risk Class + +distant human, high-wide observation + +## Reference Plan + +- Style modules: A Landscape-Led Road + B High-Angle Observation. +- Active style references: + - `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png` + - `projects/2026-06-23-在路上/intake/imgs/B.高位观察型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png` +- Visible-context roles for generation: Image 1 A road reference, Image 2 B high-observation reference, Image 3 walking/back silhouette reference. + +## Visual Requirements + +- High, wide, observational camera. +- The traveler is tiny, about 1/20 to 1/12 of frame. +- Road or route line should carry the composition. +- No triumphant vista; no tourism beauty. + +## Character Handling + +Use the walking/back slice only for silhouette, coat, backpack, and movement direction. Face is not relevant. + +## Composition / Scale + +16:9 high wide shot. The traveler may sit near a lower third or side edge, dwarfed by road and terrain. + +## Light / Color / Texture + +Cold gray, earth brown, wet dark blue-gray; documentary air and rough ground. + +## Forbidden Drift + +No drone tourism shot, no epic hero, no clean scenic highway, no bright commercial city, no cyber/neon. + +## Acceptance Checklist + +- World is much larger than traveler. +- High observation feels documentary. +- Traveler remains a small coordinate that keeps moving. +- Image status remains validation pending for parent/user confirmation. + +## Prompt Handoff + +Build the final prompt through the `gpt-image-2` Advisor / host-native workflow using A/B style references and a distant walking/back character reference. diff --git a/projects/2026-06-23-在路上/slides/s40/s40-v1-image.png b/projects/2026-06-23-在路上/slides/s40/s40-v1-image.png new file mode 100644 index 0000000..347b9fe Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s40/s40-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s40/s40-v1-prompt.md b/projects/2026-06-23-在路上/slides/s40/s40-v1-prompt.md new file mode 100644 index 0000000..9f241fb --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s40/s40-v1-prompt.md @@ -0,0 +1,60 @@ +# S-40 Slide Prompt V1 + +## Status + +- Ready for host-native image generation. +- Review mode: none. +- Target slide status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s40/s40-v1-brief.md` + +## Reference Inputs + +- Style modules: A / Landscape-Led Road; B / High-Angle City Edge observation. +- Active style references: + - `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png` + - `projects/2026-06-23-在路上/intake/imgs/B.高位观察型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png` +- Generation visible-context roles: Image 1 = A road/world reference; Image 2 = B high-observation reference; Image 3 = walking/back silhouette reference. + +## Operation Notes + +- Project-bound prompt produced through `gpt-image-2` Skill Advisor / host-native workflow. +- Garden/API expected: no. +- Actual image references must be loaded into visible context before host generation. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host has built-in `image_gen`. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Source files read: same project/task/visual-system files listed in `s38-v1-prompt.md`, plus the relevant S-40 row from `intake/阶段五-镜头清单骨架.md`. + +## Final Prompt + +Use case: photorealistic-natural +Asset type: 16:9 MV production still for slide S-40. +Input images: Image 1 is the A landscape-led road reference; Image 2 is the B high-observation reference; Image 3 is the walking/back identity reference for distant silhouette. +Primary request: Create a high, wide, photorealistic cinematic documentary still of a tiny male traveler slowly moving along a deserted rough road through a large world. +Scene/backdrop: broad cold terrain, rough muddy road or荒路, wet ruts, sparse grass or hard ground, distant emptiness. The world should feel huge and indifferent. +Subject: a very small road-worn 50-year-old male traveler, using Image 3 only for silhouette continuity: long dark coat, old backpack, old boots, walking away or across the frame. No visible face. +Style reference: use Image 1 for road/world scale and Image 2 for high observational viewpoint and system-like distance. +Composition/framing: 16:9 high wide shot, as if from a hillside, bridge, or high observation point. Road line carries the frame. Traveler about 1/20 to 1/12 of the frame, placed near a lower third or side edge. +Lighting/mood: cold overcast pre-dawn or dim gray light, documentary restraint, quiet inner space opening. +Color palette: cold gray, earth brown, dark khaki, coal black, wet blue-gray. +Textures: mud, stones, road scars, damp soil, worn coat silhouette, tiny backpack shape. +Constraints: first read must be large world and road; second read tiny traveler. No heroic stance, no scenic tourism. + +## Negative / Avoid + +Avoid drone tourism beauty, epic hero poster, centered hero backshot, clean highway, bright sunrise, bright commercial city, cyberpunk, neon, sci-fi UI, code, fashion travel photography, readable text, labels, logos, watermarks. + +## Output Intent + +Save selected host output as `projects/2026-06-23-在路上/slides/s40/s40-v1-image.png` and record status as generated / validation pending. diff --git a/projects/2026-06-23-在路上/slides/s41/s41-v1-brief.md b/projects/2026-06-23-在路上/slides/s41/s41-v1-brief.md new file mode 100644 index 0000000..e2fb0be --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s41/s41-v1-brief.md @@ -0,0 +1,70 @@ +# S-41 Slide Brief V1 + +## Status + +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Version resolved by folder scan: `s41-v1` +- Output status target: generated / validation pending + +## Source Basis + +- `intake/阶段五-镜头清单骨架.md`: S-41; night puddle close view with hinted star points. +- `slides/slides.md`: prepares final C reveal; style C; priority. +- `visual-system/visual-system.md` +- `visual-system/style/style-prompt-master.md` + +## Slide Function + +Foreshadow the soul image without fully revealing it. + +## First Read + +Real muddy water first; faint star points only discovered after the texture. + +## Risk Class + +core motif, environment-only + +## Reference Plan + +- Style module: C Muddy Starlight. +- Active style reference: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png` +- Character reference: none; no full person. +- Visible-context role for generation: Image 1 repaired C puddle reference. + +## Visual Requirements + +- Low ground-level or close puddle frame. +- Star reflection is faint, broken, and partial. +- Mud, dirty water, grit, stones, and ripple texture dominate. +- This is a preview, not the final S-46 reveal. + +## Character Handling + +None. Do not insert a full traveler or face. + +## Composition / Scale + +16:9 close or low frame. Puddle and mud occupy most of the image. + +## Light / Color / Texture + +Dark pre-dawn/night, cold whites in tiny star reflections, wet earth browns and dark blue-grays. + +## Forbidden Drift + +No clean mirror lake, no cosmic spectacle, no sunrise/sunset, no fantasy glow, no shoe advertisement. + +## Acceptance Checklist + +- Mud is real and dominant. +- Star points are visible but only hinted. +- No portrait or full-body traveler appears. +- Image status remains validation pending for parent/user confirmation. + +## Prompt Handoff + +Build the final prompt through the `gpt-image-2` Advisor / host-native workflow using the repaired C reference. diff --git a/projects/2026-06-23-在路上/slides/s41/s41-v1-image.png b/projects/2026-06-23-在路上/slides/s41/s41-v1-image.png new file mode 100644 index 0000000..fd62778 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s41/s41-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s41/s41-v1-prompt.md b/projects/2026-06-23-在路上/slides/s41/s41-v1-prompt.md new file mode 100644 index 0000000..26fe0d0 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s41/s41-v1-prompt.md @@ -0,0 +1,57 @@ +# S-41 Slide Prompt V1 + +## Status + +- Ready for host-native image generation. +- Review mode: none. +- Target slide status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s41/s41-v1-brief.md` + +## Reference Inputs + +- Style module: C / Muddy Starlight. +- Active style reference: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png` +- Character reference: none; no full person should appear. +- Generation visible-context role: Image 1 = repaired C muddy-starlight reference. + +## Operation Notes + +- Project-bound prompt produced through `gpt-image-2` Skill Advisor / host-native workflow. +- Garden/API expected: no. +- Actual image reference must be loaded into visible context before host generation. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host has built-in `image_gen`. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Source files read: same project/task/visual-system files listed in `s38-v1-prompt.md`, plus the relevant S-41 row from `intake/阶段五-镜头清单骨架.md`. + +## Final Prompt + +Use case: photorealistic-natural +Asset type: 16:9 MV production still for slide S-41. +Input images: Image 1 is the repaired C muddy-starlight reference. +Primary request: Create a dark, restrained photorealistic cinematic close view of a night puddle in real mud, with only faint broken star points hinted in the water. +Scene/backdrop: low ground-level muddy road surface, dirty water, grit, stones, uneven puddle edge, old shoe marks without showing a full person. +Subject: the muddy puddle and dark water. Use Image 1 for the accepted C texture, broken reflection, and anti-fantasy boundary. +Composition/framing: 16:9 low close frame; mud and water occupy most of the image. First read is wet mud and dirty water; second read is faint star reflection. +Lighting/mood: dark night or deep pre-dawn, minimal ambient light, tiny cold-white star points broken by ripples and mud. Quiet foreshadowing, not the full reveal. +Color palette: wet dark blue-gray, earth brown, coal black, tiny cold-white star points. +Textures: real wet mud, gritty puddle edge, ripples, stones, dirty water surface, broken shoe marks. +Constraints: star reflection must be visible but partial and restrained. No full-body traveler and no face. No text. + +## Negative / Avoid + +Avoid sunrise, sunset, warm horizon, clean mirror lake, pretty postcard reflection, fantasy portal, magic glow, cosmic spectacle, shoe advertisement, full-body traveler, portrait, face, only mud without stars, only stars without mud, readable text, logos, watermarks. + +## Output Intent + +Save selected host output as `projects/2026-06-23-在路上/slides/s41/s41-v1-image.png` and record status as generated / validation pending. diff --git a/projects/2026-06-23-在路上/slides/s42/s42-v1-brief.md b/projects/2026-06-23-在路上/slides/s42/s42-v1-brief.md new file mode 100644 index 0000000..cdc0b37 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s42/s42-v1-brief.md @@ -0,0 +1,73 @@ +# S-42 Slide Brief V1 + +## Status + +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Version resolved by folder scan: `s42-v1` +- Output status target: generated / validation pending + +## Source Basis + +- `intake/阶段五-镜头清单骨架.md`: S-42; far small fire and near weak starlight puddle, traveler passes between. +- `slides/slides.md`: connects D and C quietly; style C / D. +- `visual-system/visual-system.md` +- `visual-system/style/style-prompt-master.md` +- `visual-system/characters/character-reference-registry.md` + +## Slide Function + +Connect fire and starlight without explaining them. + +## First Read + +Night road/mud and near puddle first; far fire second; traveler passing between them third. + +## Risk Class + +distant human, C/D motif transition + +## Reference Plan + +- Style modules: C Muddy Starlight + D Night Road Fire. +- Active style references: + - `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png` + - `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png` +- Visible-context roles for generation: Image 1 C puddle reference, Image 2 D fire reference, Image 3 walking side identity reference. + +## Visual Requirements + +- Near foreground puddle carries weak broken starlight. +- Small fire is far away, not a destination. +- Traveler passes between puddle and fire as a small side/back silhouette. +- Keep the frame quiet and transitional. + +## Character Handling + +Use walking side slice for coat, backpack, boots, and passing movement. No face. + +## Composition / Scale + +Wide 16:9 frame with near low puddle foreground and far fire in background; traveler small, about 1/10 to 1/7 of frame. + +## Light / Color / Texture + +Cold muddy night with tiny cold star reflection and restrained warm-orange distant fire. + +## Forbidden Drift + +No camping, reunion, fantasy bridge, glowing portal, large person, or explanatory symbolism. + +## Acceptance Checklist + +- C and D coexist without spectacle. +- Traveler is moving through, not arriving. +- Fire remains far and small. +- Image status remains validation pending for parent/user confirmation. + +## Prompt Handoff + +Build the final prompt through the `gpt-image-2` Advisor / host-native workflow using C/D style references and a walking side character reference. diff --git a/projects/2026-06-23-在路上/slides/s42/s42-v1-image.png b/projects/2026-06-23-在路上/slides/s42/s42-v1-image.png new file mode 100644 index 0000000..0dd8203 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s42/s42-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s42/s42-v1-prompt.md b/projects/2026-06-23-在路上/slides/s42/s42-v1-prompt.md new file mode 100644 index 0000000..d3af20f --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s42/s42-v1-prompt.md @@ -0,0 +1,60 @@ +# S-42 Slide Prompt V1 + +## Status + +- Ready for host-native image generation. +- Review mode: none. +- Target slide status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s42/s42-v1-brief.md` + +## Reference Inputs + +- Style modules: C / Muddy Starlight; D / Night Road Fire. +- Active style references: + - `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png` + - `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png` +- Generation visible-context roles: Image 1 = repaired C puddle reference; Image 2 = D fire reference; Image 3 = walking side identity reference. + +## Operation Notes + +- Project-bound prompt produced through `gpt-image-2` Skill Advisor / host-native workflow. +- Garden/API expected: no. +- Actual image references must be loaded into visible context before host generation. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host has built-in `image_gen`. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Source files read: same project/task/visual-system files listed in `s38-v1-prompt.md`, plus the relevant S-42 row from `intake/阶段五-镜头清单骨架.md`. + +## Final Prompt + +Use case: photorealistic-natural +Asset type: 16:9 MV production still for slide S-42. +Input images: Image 1 is the repaired C muddy-starlight reference; Image 2 is the D night-road fire reference; Image 3 is the walking side identity reference. +Primary request: Create a restrained photorealistic cinematic night-road still where a near muddy puddle contains weak broken starlight, a far small roadside fire glows in the distance, and the road-worn traveler passes quietly between them. +Scene/backdrop: dark wet wilderness road, foreground muddy puddle with gritty dirty water, far road darkness, one small distant fire point, sparse stones and wet ruts. +Subject: small 50-year-old male traveler passing side/back across the road, long dark coat, old backpack, old boots, steady tired movement. Use Image 3 for silhouette and outfit continuity; no face. +Style references: use Image 1 for real mud, dark water, and broken star reflection; use Image 2 for restrained small firelight and anti-camping mood. +Composition/framing: wide 16:9 frame. Foreground puddle low in frame, far fire small in background, traveler between them as a small moving coordinate about 1/10 to 1/7 of frame. Road continues beyond all elements. +Lighting/mood: cold blue-gray night, tiny cold-white star reflection, small warm-orange distant fire. Quiet transition, not revelation. +Color palette: wet dark blue-gray, earth brown, coal black, restrained orange, tiny cold-white points. +Textures: wet mud, rippled puddle, stones, coat fabric, old backpack, boot silhouette, road grit. +Constraints: fire is a signal, not a campsite or destination; star reflection is weak and broken; traveler continues moving. + +## Negative / Avoid + +Avoid camping setup, tent, chairs, cookware, cozy camp, group, reunion, victory, fantasy portal, magic glow, cosmic spectacle, large centered person, portrait, bright sunrise, bright sunset, high saturation, readable text, labels, logos, watermarks. + +## Output Intent + +Save selected host output as `projects/2026-06-23-在路上/slides/s42/s42-v1-image.png` and record status as generated / validation pending. diff --git a/projects/2026-06-23-在路上/slides/s43/s43-v1-brief.md b/projects/2026-06-23-在路上/slides/s43/s43-v1-brief.md new file mode 100644 index 0000000..33e1f4a --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s43/s43-v1-brief.md @@ -0,0 +1,70 @@ +# S-43 Slide Brief V1 + +## Status + +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Version resolved by folder scan: `s43-v1` +- Output status target: generated / validation pending + +## Source Basis + +- `intake/阶段五-镜头清单骨架.md`: S-43; traveler stops before a puddle, almost still. +- `slides/slides.md`: pause before the soul reveal; style C. +- `visual-system/visual-system.md` +- `visual-system/style/style-prompt-master.md` +- `visual-system/characters/character-reference-registry.md` + +## Slide Function + +Create a held breath before the final four lines. + +## First Read + +Still night road and puddle first; stopped back-view traveler second. + +## Risk Class + +medium human, C motif pause + +## Reference Plan + +- Style module: C Muddy Starlight. +- Active style reference: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v2-from-anchor.png` +- Visible-context roles for generation: Image 1 C puddle reference, Image 2 back identity reference. + +## Visual Requirements + +- Traveler stops before a puddle, seen from back or back-side. +- Almost no motion; the frame feels still and waiting. +- Puddle has dark water and maybe tiny star hints, not full reveal. + +## Character Handling + +Use back full-body slice for coat, backpack, shoulder line, and non-hunched posture. + +## Composition / Scale + +Medium-wide 16:9 frame. Traveler about 1/8 to 1/6 of frame, not centered heroically. + +## Light / Color / Texture + +Cold night or deep pre-dawn, wet mud, dark water, minimal cold highlights. + +## Forbidden Drift + +No portrait, no face, no dramatic kneeling, no magical revelation, no clean reflection lake. + +## Acceptance Checklist + +- Stillness is the main action. +- Traveler is restrained and back-view. +- Puddle remains real and low. +- Image status remains validation pending for parent/user confirmation. + +## Prompt Handoff + +Build the final prompt through the `gpt-image-2` Advisor / host-native workflow using C style and back-character reference. diff --git a/projects/2026-06-23-在路上/slides/s43/s43-v1-image.png b/projects/2026-06-23-在路上/slides/s43/s43-v1-image.png new file mode 100644 index 0000000..d291e30 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s43/s43-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s43/s43-v1-prompt.md b/projects/2026-06-23-在路上/slides/s43/s43-v1-prompt.md new file mode 100644 index 0000000..4976213 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s43/s43-v1-prompt.md @@ -0,0 +1,58 @@ +# S-43 Slide Prompt V1 + +## Status + +- Ready for host-native image generation. +- Review mode: none. +- Target slide status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s43/s43-v1-brief.md` + +## Reference Inputs + +- Style module: C / Muddy Starlight. +- Active style reference: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v2-from-anchor.png` +- Generation visible-context roles: Image 1 = repaired C puddle reference; Image 2 = back identity reference. + +## Operation Notes + +- Project-bound prompt produced through `gpt-image-2` Skill Advisor / host-native workflow. +- Garden/API expected: no. +- Actual image references must be loaded into visible context before host generation. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host has built-in `image_gen`. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Source files read: same project/task/visual-system files listed in `s38-v1-prompt.md`, plus the relevant S-43 row from `intake/阶段五-镜头清单骨架.md`. + +## Final Prompt + +Use case: photorealistic-natural +Asset type: 16:9 MV production still for slide S-43. +Input images: Image 1 is the repaired C muddy-starlight reference; Image 2 is the back identity reference for the traveler. +Primary request: Create a nearly still photorealistic cinematic documentary night-road frame: the traveler has stopped before a muddy puddle, waiting in silence before the final reveal. +Scene/backdrop: dark wet muddy road, shallow puddle in front of the traveler, dirty water, stones, low night air, sparse terrain. +Subject: a 50-year-old male traveler seen from back or back-side, long dark coat, old backpack, old boots, upright but tired. Use Image 2 for back silhouette, backpack, coat length, and posture. No face. +Style reference: use Image 1 for real wet mud, dark water, broken faint star hints, and anti-fantasy boundary. +Composition/framing: medium-wide 16:9 frame. Puddle and wet ground first; stopped traveler second, about 1/8 to 1/6 of frame, slightly off-center. The frame should feel almost motionless. +Lighting/mood: cold night or deep pre-dawn, minimal ambient blue-gray light, maybe tiny cold star hints in the puddle, no full revelation yet. +Color palette: wet dark blue-gray, earth brown, coal black, muted dark khaki, tiny cold-white points if visible. +Textures: wet mud, puddle surface, stones, old boot edges, coat fabric, backpack seams. +Constraints: held breath before the soul image; documentary reality; no magical glow; no portrait. + +## Negative / Avoid + +Avoid front-facing portrait, face detail, heroic centered stance, kneeling melodrama, clean mirror lake, fantasy portal, magic glow, cosmic spectacle, sunrise, sunset, shoe advertisement, bright colors, readable text, labels, logos, watermarks. + +## Output Intent + +Save selected host output as `projects/2026-06-23-在路上/slides/s43/s43-v1-image.png` and record status as generated / validation pending. diff --git a/projects/2026-06-23-在路上/slides/s44/s44-v1-brief.md b/projects/2026-06-23-在路上/slides/s44/s44-v1-brief.md new file mode 100644 index 0000000..04cd592 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s44/s44-v1-brief.md @@ -0,0 +1,72 @@ +# S-44 Slide Brief V1 + +## Status + +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Version resolved by folder scan: `s44-v1` +- Output status target: generated / validation pending + +## Source Basis + +- `intake/阶段五-镜头清单骨架.md`: S-44, lyric "向前走"; traveler keeps walking and does not look back. +- `slides/slides.md`: action replaces slogan; style A / D. +- `visual-system/visual-system.md` +- `visual-system/style/style-prompt-master.md` +- `visual-system/characters/character-reference-registry.md` + +## Slide Function + +Show forward motion without turning it into a slogan or victory image. + +## First Read + +Night road and forward direction first; tiny back-view traveler continuing second. + +## Risk Class + +distant human, continuation action + +## Reference Plan + +- Style modules: A Landscape-Led Road + D Night Road Fire warmth if needed. +- Active style references: + - `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png` + - `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png` +- Visible-context roles for generation: Image 1 A road reference, Image 2 D night-fire reference, Image 3 walking back identity reference. + +## Visual Requirements + +- Traveler continues forward from one side of frame or along the road. +- No looking back, no face. +- Road stays dominant; fire/warmth may be faint and secondary. + +## Character Handling + +Use walking back slice for no-looking-back posture, coat, backpack, boots, and forward motion. + +## Composition / Scale + +Wide 16:9 frame. Traveler small, about 1/12 to 1/8 of frame. + +## Light / Color / Texture + +Cold night/pre-dawn road with restrained warm point if used; wet mud and worn road texture. + +## Forbidden Drift + +No heroic silhouette, no motivational poster, no arms-open pose, no sunset road victory. + +## Acceptance Checklist + +- Action is simply "keep walking." +- Traveler does not look back. +- Road is still the protagonist. +- Image status remains validation pending for parent/user confirmation. + +## Prompt Handoff + +Build the final prompt through the `gpt-image-2` Advisor / host-native workflow using A/D style references and walking back character reference. diff --git a/projects/2026-06-23-在路上/slides/s44/s44-v1-image.png b/projects/2026-06-23-在路上/slides/s44/s44-v1-image.png new file mode 100644 index 0000000..4018c74 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s44/s44-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s44/s44-v1-prompt.md b/projects/2026-06-23-在路上/slides/s44/s44-v1-prompt.md new file mode 100644 index 0000000..2dc0984 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s44/s44-v1-prompt.md @@ -0,0 +1,60 @@ +# S-44 Slide Prompt V1 + +## Status + +- Ready for host-native image generation. +- Review mode: none. +- Target slide status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s44/s44-v1-brief.md` + +## Reference Inputs + +- Style modules: A / Landscape-Led Road; D / Night Road Fire. +- Active style references: + - `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png` + - `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png` +- Generation visible-context roles: Image 1 = A road/world reference; Image 2 = D night-fire reference; Image 3 = walking back identity reference. + +## Operation Notes + +- Project-bound prompt produced through `gpt-image-2` Skill Advisor / host-native workflow. +- Garden/API expected: no. +- Actual image references must be loaded into visible context before host generation. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host has built-in `image_gen`. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Source files read: same project/task/visual-system files listed in `s38-v1-prompt.md`, plus the relevant S-44 row from `intake/阶段五-镜头清单骨架.md`. + +## Final Prompt + +Use case: photorealistic-natural +Asset type: 16:9 MV production still for slide S-44. +Input images: Image 1 is the A landscape-led road reference; Image 2 is the D night-road fire reference; Image 3 is the walking back identity reference. +Primary request: Create a restrained photorealistic cinematic documentary still for the lyric "walk forward": the road-worn traveler keeps walking away along the night road and does not look back. +Scene/backdrop: cold muddy road at night or deep pre-dawn, wet ruts, stones, sparse dark terrain, road extending forward into darkness. A very small warm fire point may exist far behind or off to the side, but it is not the destination. +Subject: a 50-year-old male traveler seen only from the back, walking forward, long dark coat, old backpack, old boots. Use Image 3 for walking-away silhouette and outfit continuity. No face, no turning head. +Style references: use Image 1 for world-before-person road scale; use Image 2 only for restrained night warmth and anti-camping boundary. +Composition/framing: wide 16:9 frame. Road dominates. Traveler small, about 1/12 to 1/8 of frame, moving from one side or along the road into the distance. +Lighting/mood: cold wet blue-gray night with restrained warmth if any. The feeling is continuation, not triumph. +Color palette: low saturation cold gray, earth brown, coal black, wet dark blue-gray, tiny warm orange if used. +Textures: wet mud, road scars, old coat, backpack, worn boots, damp air. +Constraints: show action instead of slogan. The traveler must not look back. No text in image. + +## Negative / Avoid + +Avoid motivational poster, heroic silhouette, arms-open pose, centered victory stance, sunset road, travel ad, camping scene, group reunion, large fire, bright horizon, fashion photography, readable text, labels, logos, watermarks, UI. + +## Output Intent + +Save selected host output as `projects/2026-06-23-在路上/slides/s44/s44-v1-image.png` and record status as generated / validation pending. diff --git a/projects/2026-06-23-在路上/slides/s45/s45-v1-brief.md b/projects/2026-06-23-在路上/slides/s45/s45-v1-brief.md new file mode 100644 index 0000000..9435106 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s45/s45-v1-brief.md @@ -0,0 +1,71 @@ +# S-45 Slide Brief V1 + +## Status + +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Version resolved by folder scan: `s45-v1` +- Output status target: generated / validation pending + +## Source Basis + +- `intake/阶段五-镜头清单骨架.md`: S-45, lyric "别回头"; back view passes old footprints without looking back. +- `slides/slides.md`: closure into "continue"; style A / motif. +- `visual-system/visual-system.md` +- `visual-system/style/style-prompt-master.md` +- `visual-system/characters/character-reference-registry.md` + +## Slide Function + +Compress farewell into a physical action: old footprints remain behind; the traveler continues. + +## First Read + +Muddy footprints and old road trace first; back-view traveler moving past them second. + +## Risk Class + +medium/distant human, motif detail + +## Reference Plan + +- Style module: A Landscape-Led Road. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png` +- Visible-context roles for generation: Image 1 A road reference, Image 2 walking back identity reference. + +## Visual Requirements + +- Back view from behind the traveler. +- Old footprints visible in wet mud behind or beside him. +- He does not turn back. +- Frame leads into the coming C reveal; keep mud real. + +## Character Handling + +Use walking back slice for posture, coat/backpack silhouette, boots, and no-looking-back direction. + +## Composition / Scale + +Medium-wide or wide 16:9 frame. Footprints occupy foreground/midground; traveler about 1/10 to 1/7 of frame. + +## Light / Color / Texture + +Cold dark road or pre-dawn, damp mud, worn boot prints, restrained documentary light. + +## Forbidden Drift + +No sentimental portrait, no readable text, no clean symbolic footprints, no bright inspirational road. + +## Acceptance Checklist + +- Footprints are old and physical, not graphic symbols. +- Traveler continues away without turning. +- Mud and road remain real. +- Image status remains validation pending for parent/user confirmation. + +## Prompt Handoff + +Build the final prompt through the `gpt-image-2` Advisor / host-native workflow using A style and walking back character reference. diff --git a/projects/2026-06-23-在路上/slides/s45/s45-v1-image.png b/projects/2026-06-23-在路上/slides/s45/s45-v1-image.png new file mode 100644 index 0000000..45b42e7 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s45/s45-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s45/s45-v1-prompt.md b/projects/2026-06-23-在路上/slides/s45/s45-v1-prompt.md new file mode 100644 index 0000000..9425a65 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s45/s45-v1-prompt.md @@ -0,0 +1,58 @@ +# S-45 Slide Prompt V1 + +## Status + +- Ready for host-native image generation. +- Review mode: none. +- Target slide status after generation: generated / validation pending. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s45/s45-v1-brief.md` + +## Reference Inputs + +- Style module: A / Landscape-Led Road. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png` +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png` +- Generation visible-context roles: Image 1 = A road/world reference; Image 2 = walking back identity reference. + +## Operation Notes + +- Project-bound prompt produced through `gpt-image-2` Skill Advisor / host-native workflow. +- Garden/API expected: no. +- Actual image references must be loaded into visible context before host generation. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json` +- Mode result: `B-or-C`; interpreted as Mode B because the host has built-in `image_gen`. +- Skill/template files read: + - `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\prompt-writing.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md` + - `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md` +- Source files read: same project/task/visual-system files listed in `s38-v1-prompt.md`, plus the relevant S-45 row from `intake/阶段五-镜头清单骨架.md`. + +## Final Prompt + +Use case: photorealistic-natural +Asset type: 16:9 MV production still for slide S-45. +Input images: Image 1 is the A landscape-led muddy road reference; Image 2 is the walking back identity reference. +Primary request: Create a restrained photorealistic cinematic documentary still from behind the traveler as he passes old muddy footprints and keeps walking without looking back. +Scene/backdrop: cold muddy road at night or deep pre-dawn, wet ruts, stones, old boot prints in mud, sparse rough ground, road continuing forward. +Subject: a 50-year-old male traveler seen from the back, long dark coat, old backpack, old boots, walking away. Use Image 2 for back silhouette, backpack, coat, boots, and no-looking-back posture. No face. +Style reference: use Image 1 for landscape-led road, mud texture, and world-before-person composition. +Composition/framing: medium-wide or wide 16:9 frame. Old footprints occupy foreground or midground; traveler about 1/10 to 1/7 of frame, moving away from the footprints and not turning his head. +Lighting/mood: cold gray-blue night or pre-dawn, restrained documentary light, quiet farewell compressed into action. +Color palette: cold gray, earth brown, wet dark blue-gray, coal black, muted dark khaki. +Textures: wet mud, footprint edges, water-filled ruts, worn boots, rough coat fabric, old backpack. +Constraints: footprints must feel physical and real, not symbolic graphics. The traveler must continue forward and must not look back. No text. + +## Negative / Avoid + +Avoid sentimental portrait, front face, turned head, motivational poster, clean symbolic footprints, bright inspirational road, sunrise, sunset, travel ad, fashion photography, heroic stance, readable text, labels, logos, watermarks, UI. + +## Output Intent + +Save selected host output as `projects/2026-06-23-在路上/slides/s45/s45-v1-image.png` and record status as generated / validation pending. diff --git a/projects/2026-06-23-在路上/slides/s48/s48-v1-brief.md b/projects/2026-06-23-在路上/slides/s48/s48-v1-brief.md new file mode 100644 index 0000000..a13cee5 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s48/s48-v1-brief.md @@ -0,0 +1,78 @@ +# S-48 Slide Brief V1 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Version: `s48-v1`. +- Overwrite allowed: no. + +## Source Basis + +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md`: S-48 row. +- `projects/2026-06-23-在路上/slides/slides.md`: S-48 execution row. +- `projects/2026-06-23-在路上/visual-system/visual-system.md`. +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md`. +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md`. + +## Slide Function + +Second and third weak fire points appear in the far darkness. Other travelers become visible only as small fire signals, not as people or a group. + +## First Read + +Night road and large darkness first, then the near small fire, then two or three far weak fire points. + +## Risk Class + +- Core motif. +- Distant human / mostly landscape. + +## Reference Plan + +- Style module: D, Night Road Fire. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png`. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-back-reference-crop-tight.png` only for tiny silhouette continuity if the traveler is visible. +- Visible-context roles for host generation: Image 1 style reference, Image 2 distant back/walking character reference. + +## Visual Requirements + +- 16:9 photorealistic cinematic documentary road-film still. +- Dark wet wilderness road at night or deep pre-dawn. +- Near small roadside fire remains restrained. +- Far fire points are sparse and weak, not bright lamps or a camp. +- No clear distant people; the others are suggested only by fire. + +## Character Handling + +The traveler may be small, seated or standing near the near fire, side/back only. Identity detail is not the point; coat and backpack silhouette should stay consistent. + +## Composition / Scale + +- Very wide or wide frame. +- World and road dominate. +- Traveler no larger than about 1/12 of frame if present. +- Far fire points occupy only tiny parts of the horizon / road distance. + +## Light / Color / Texture + +- Low saturation: cold gray, earth brown, coal black, wet dark blue-gray. +- Warmth only from small orange firelight. +- Mud, puddles, stones, rough wet road texture must remain real. + +## Forbidden Drift + +No camping ad, no reunion, no visible group, no party, no victory, no large theatrical fire, no tent, no cookware, no warm sunset, no scenic tourism, no fantasy glow, no text, no logo, no watermark. + +## Acceptance Checklist + +- First read is darkness / road, not person. +- Far fires suggest others without showing a gathering. +- Fire is a signal, not a destination. +- Ending remains restrained and unfinished. +- No `s48-v1-review.md` is created. + +## Prompt Handoff + +Use `gpt-image-2` Advisor / host-native workflow. Load Image 1 as D style reference and Image 2 as distant back/walking character reference before host generation. Save output to `projects/2026-06-23-在路上/slides/s48/s48-v1-image.png`. diff --git a/projects/2026-06-23-在路上/slides/s48/s48-v1-image.png b/projects/2026-06-23-在路上/slides/s48/s48-v1-image.png new file mode 100644 index 0000000..dde04d0 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s48/s48-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s48/s48-v1-prompt.md b/projects/2026-06-23-在路上/slides/s48/s48-v1-prompt.md new file mode 100644 index 0000000..774dc6c --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s48/s48-v1-prompt.md @@ -0,0 +1,55 @@ +# S-48 Slide Prompt V1 + +## Status + +- Line: production-light. +- Review mode: none. +- Version: `s48-v1`. +- Output target: `projects/2026-06-23-在路上/slides/s48/s48-v1-image.png`. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s48/s48-v1-brief.md`. + +## Reference Inputs + +- Style module: D / Night Road Fire. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png`. +- Reference role: style, composition, mood, restrained firelight, distant fire-point grammar. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-back-reference-crop-tight.png`. +- Character role: distant silhouette continuity only; no close identity rendering required. +- Visible context loaded before generation: Image 1 = D style reference; Image 2 = distant back/walking character reference. + +## Operation Notes + +- Resolved version by scanning `slides/s48/`; no existing versioned files were found. +- Host-native generation is authorized by task mode `prompt and generate`. +- Review file is intentionally absent because review mode is `none`. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json`. +- Mode result: `B-or-C`; interpreted as Mode B because host `image_gen` is available. +- Garden/API execution expected: no. +- Skill files read: `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md`. +- Template/reference files read: `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md`; `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md`. +- Project source files read: `AGENTS.md`; `VIDEO_WORKBENCH.md`; `docs/workflows/project-lifecycle.md`; `docs/workflows/slide-task-orchestration.md`; `docs/workflows/slide-task-templates.md`; `handoff/2026-06-25-on-the-road-formal-production-handoff.md`; `project.md`; `execution-plan.md`; `slides/slides.md`; task plan/status; `visual-system.md`; `style-prompt-master.md`; `character-reference-registry.md`; S-48 intake row. + +## Final Prompt + +Use Image 1 as the night-road-fire style and mood reference. Use Image 2 only for the tiny traveler's back/coat/backpack silhouette if a traveler is visible; do not make the person a portrait subject. + +Create a 16:9 photorealistic cinematic documentary road-film still. A dark wet wilderness road at night or deep pre-dawn dominates the frame. The road, mud, puddles, stones, and surrounding darkness are the first read. A small restrained roadside fire glows near the lower side of the frame; the traveler is very small beside it or barely visible as a side/back silhouette. In the far darkness along the road, two or three separate weak orange fire points appear for the first time, tiny and distant, suggesting other travelers without showing any people or group. + +Composition: very wide road-film landscape, environment first, road second, near fire third, far fire points fourth, traveler last. Keep the traveler no larger than about 1/12 of the frame if present. Keep distant fire points sparse, low, and weak; they must read as fire signals, not lamps or a campsite. Preserve documentary realism and rough wet road texture. + +Style: real documentary texture with restrained poetic realism, low saturation, cold gray, earth brown, dark khaki, coal black, wet dark blue-gray. Warmth exists only as small local orange firelight and tiny distant fire points. The emotion is less alone but unfinished, with no arrival. + +## Negative / Avoid + +Avoid camping advertisement, cozy campsite, tent, chairs, cookware, picnic setup, visible group, reunion, party, victory, heroic pose, large theatrical campfire, warm sunset, scenic tourism, fantasy magic, glowing portal, neon, cyberpunk, UI, readable text, labels, logos, watermarks. + +## Output Intent + +- Save selected host-generated image to `projects/2026-06-23-在路上/slides/s48/s48-v1-image.png`. +- Record as `production-pass / final production direct-child` in `slides/slides.md` after generation succeeds. diff --git a/projects/2026-06-23-在路上/slides/s49/s49-v1-brief.md b/projects/2026-06-23-在路上/slides/s49/s49-v1-brief.md new file mode 100644 index 0000000..1961cc9 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s49/s49-v1-brief.md @@ -0,0 +1,77 @@ +# S-49 Slide Brief V1 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Version: `s49-v1`. +- Overwrite allowed: no. + +## Source Basis + +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md`: S-49 row. +- `projects/2026-06-23-在路上/slides/slides.md`: S-49 execution row. +- `projects/2026-06-23-在路上/visual-system/visual-system.md`. +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md`. +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md`. + +## Slide Function + +The traveler quietly notices the distant fires, confirms that others exist, but does not run, wave, or approach. + +## First Read + +Side/back pause in darkness, then line of sight toward distant weak fire points. + +## Risk Class + +- Distant to medium human. +- Core ending restraint. + +## Reference Plan + +- Style module: D, Night Road Fire. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png`. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-side-reference-crop-tight.png`. +- Visible-context roles for host generation: Image 1 style reference, Image 3 side/walking character identity and posture reference. + +## Visual Requirements + +- 16:9 photorealistic cinematic documentary road-film still. +- Traveler seen from side/back or three-quarter back, quiet and still. +- Far fires visible but unreachable. +- No running, waving, reunion, or clear group. + +## Character Handling + +Use the side/walking character reference for coat, backpack, glasses silhouette/profile, tired dignity, and road-worn posture. The traveler should be readable but still secondary to the night road. + +## Composition / Scale + +- Medium-wide to wide side/back frame. +- Traveler about 1/8 to 1/6 of frame. +- Traveler is stopped, shoulders calm, head turned toward far fire points. +- Road continues between traveler and far fire; distance remains real. + +## Light / Color / Texture + +- Cold dark night / deep pre-dawn environment. +- Small warm firelight may edge the coat and face profile. +- Wet mud, puddles, road ruts, stones, and rough ground remain documentary. + +## Forbidden Drift + +No waving, no raised arms, no running, no reunion, no group, no hero pose, no romantic backlit portrait, no camping scene, no large fire, no victorious smile, no text, no logo, no watermark. + +## Acceptance Checklist + +- The traveler confirms the far fire but does not approach. +- Mood is quiet recognition, not celebration. +- Character silhouette, coat, backpack, and profile stay stable. +- Night road and distance remain larger than the person. +- No `s49-v1-review.md` is created. + +## Prompt Handoff + +Use `gpt-image-2` Advisor / host-native workflow. Load Image 1 as D style reference and Image 3 as side/walking character reference before host generation. Save output to `projects/2026-06-23-在路上/slides/s49/s49-v1-image.png`. diff --git a/projects/2026-06-23-在路上/slides/s49/s49-v1-image.png b/projects/2026-06-23-在路上/slides/s49/s49-v1-image.png new file mode 100644 index 0000000..c92d857 Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s49/s49-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s49/s49-v1-prompt.md b/projects/2026-06-23-在路上/slides/s49/s49-v1-prompt.md new file mode 100644 index 0000000..635fb31 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s49/s49-v1-prompt.md @@ -0,0 +1,55 @@ +# S-49 Slide Prompt V1 + +## Status + +- Line: production-light. +- Review mode: none. +- Version: `s49-v1`. +- Output target: `projects/2026-06-23-在路上/slides/s49/s49-v1-image.png`. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s49/s49-v1-brief.md`. + +## Reference Inputs + +- Style module: D / Night Road Fire. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png`. +- Reference role: style, composition, mood, restrained firelight, far fire-point distance. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-side-reference-crop-tight.png`. +- Character role: side/back posture, coat, backpack, glasses/profile silhouette, tired dignity. +- Visible context loaded before generation: Image 1 = D style reference; Image 3 = side/walking character reference. + +## Operation Notes + +- Resolved version by scanning `slides/s49/`; no existing versioned files were found. +- Host-native generation is authorized by task mode `prompt and generate`. +- Review file is intentionally absent because review mode is `none`. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json`. +- Mode result: `B-or-C`; interpreted as Mode B because host `image_gen` is available. +- Garden/API execution expected: no. +- Skill files read: `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md`. +- Template/reference files read: `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md`; `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md`. +- Project source files read: `AGENTS.md`; `VIDEO_WORKBENCH.md`; `docs/workflows/project-lifecycle.md`; `docs/workflows/slide-task-orchestration.md`; `docs/workflows/slide-task-templates.md`; `handoff/2026-06-25-on-the-road-formal-production-handoff.md`; `project.md`; `execution-plan.md`; `slides/slides.md`; task plan/status; `visual-system.md`; `style-prompt-master.md`; `character-reference-registry.md`; S-49 intake row. + +## Final Prompt + +Use Image 1 as the night-road-fire style and mood reference. Use Image 3 as the traveler identity/posture reference: a 50-year-old road-worn man with dark slightly messy hair, stubble, black rectangular glasses, long dark coat, old backpack, old boots, tired but dignified side/back posture. + +Create a 16:9 photorealistic cinematic documentary road-film still. On a dark wet wilderness road at night or deep pre-dawn, the traveler stands still in a side/back or three-quarter-back view, quietly looking toward two or three weak distant fire points far down the road. He does not move toward them. He does not wave. His body is calm and upright, tired but capable, as if he has only confirmed that others exist somewhere ahead. + +Composition: medium-wide to wide road-film frame. The night road and distance dominate. Traveler about 1/8 to 1/6 of the frame, placed off-center near the foreground or midground edge. A small near fire or faint local firelight may rim his coat and backpack, but the far fires remain tiny and separated in the darkness. The road continues between him and the far fire points; no reunion or arrival. + +Style: real documentary texture with restrained poetic realism, low saturation, cold gray, earth brown, dark khaki, coal black, wet dark blue-gray. Warmth appears only as small orange firelight. Wet mud, puddles, road ruts, stones, worn coat, old backpack, and boots should feel physically real. + +## Negative / Avoid + +Avoid running, waving, raised arms, group, reunion, party, victory, heroic centered stance, romantic portrait, fashion photography, cozy campsite, tent, camping chairs, large theatrical campfire, warm sunset, scenic tourism, fantasy glow, neon, cyberpunk, UI, readable text, labels, logos, watermarks. + +## Output Intent + +- Save selected host-generated image to `projects/2026-06-23-在路上/slides/s49/s49-v1-image.png`. +- Record as `production-pass / final production direct-child` in `slides/slides.md` after generation succeeds. diff --git a/projects/2026-06-23-在路上/slides/s50/s50-v1-brief.md b/projects/2026-06-23-在路上/slides/s50/s50-v1-brief.md new file mode 100644 index 0000000..d1800d0 --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s50/s50-v1-brief.md @@ -0,0 +1,78 @@ +# S-50 Slide Brief V1 + +## Status + +- Line: production-light. +- Running mode: prompt and generate. +- Review mode: none. +- Version: `s50-v1`. +- Overwrite allowed: no. + +## Source Basis + +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md`: S-50 row. +- `projects/2026-06-23-在路上/slides/slides.md`: S-50 execution row. +- `projects/2026-06-23-在路上/visual-system/visual-system.md`. +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md`. +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md`. + +## Slide Function + +The traveler leaves his small fire, shoulders the backpack, and continues down the night road. Fire is not the destination; it is the reason to keep going. + +## First Read + +Back view walking away from the small fire into the continuing night road. + +## Risk Class + +- Distant human. +- Core ending action. + +## Reference Plan + +- Style module: D, Night Road Fire. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png`. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`. +- Visible-context roles for host generation: Image 1 style reference, Image 4 walking-back character reference. + +## Visual Requirements + +- 16:9 photorealistic cinematic documentary road-film still. +- Back view of traveler walking away from his own small fire. +- Backpack and long coat visible. +- Road continues ahead toward darkness and weak far fire points. +- No arrival, no final group, no victory. + +## Character Handling + +Use the walking-back character slice for backpack, coat length, shoulder line, stride, and boots. The traveler must remain upright and capable, not collapsed. + +## Composition / Scale + +- Wide road-film frame. +- Traveler about 1/10 to 1/7 of frame, back view, leaving the near fire behind. +- Fire about 3% to 6% of frame, behind or side-behind the traveler. +- Road must continue ahead and remain larger than the person. + +## Light / Color / Texture + +- Cold night / deep pre-dawn darkness. +- Small warm firelight behind the traveler, weak edge glow only. +- Wet mud, puddles, road ruts, stones, worn boots, and coat texture remain real. + +## Forbidden Drift + +No campsite comfort, no final arrival, no group, no farewell wave, no heroic ending, no broad sunrise, no bright path of destiny, no fantasy glow, no text, no logo, no watermark. + +## Acceptance Checklist + +- The image says "continues" rather than "arrives". +- The local fire is behind or beside him, not the destination. +- Traveler identity reads through back silhouette, backpack, coat, and stride. +- Road and darkness remain larger than the person. +- No `s50-v1-review.md` is created. + +## Prompt Handoff + +Use `gpt-image-2` Advisor / host-native workflow. Load Image 1 as D style reference and Image 4 as walking-back character reference before host generation. Save output to `projects/2026-06-23-在路上/slides/s50/s50-v1-image.png`. diff --git a/projects/2026-06-23-在路上/slides/s50/s50-v1-image.png b/projects/2026-06-23-在路上/slides/s50/s50-v1-image.png new file mode 100644 index 0000000..a21180d Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s50/s50-v1-image.png differ diff --git a/projects/2026-06-23-在路上/slides/s50/s50-v1-prompt.md b/projects/2026-06-23-在路上/slides/s50/s50-v1-prompt.md new file mode 100644 index 0000000..6afa3cf --- /dev/null +++ b/projects/2026-06-23-在路上/slides/s50/s50-v1-prompt.md @@ -0,0 +1,55 @@ +# S-50 Slide Prompt V1 + +## Status + +- Line: production-light. +- Review mode: none. +- Version: `s50-v1`. +- Output target: `projects/2026-06-23-在路上/slides/s50/s50-v1-image.png`. + +## Based On Brief + +- `projects/2026-06-23-在路上/slides/s50/s50-v1-brief.md`. + +## Reference Inputs + +- Style module: D / Night Road Fire. +- Active style reference: `projects/2026-06-23-在路上/intake/imgs/D.夜路篝火型蓝图.png`. +- Reference role: style, composition, mood, restrained firelight, continuing road. +- Character reference: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`. +- Character role: walking-back silhouette, backpack, long coat, shoulder line, stride, boots. +- Visible context loaded before generation: Image 1 = D style reference; Image 4 = walking-back character reference. + +## Operation Notes + +- Resolved version by scanning `slides/s50/`; no existing versioned files were found. +- Host-native generation is authorized by task mode `prompt and generate`. +- Review file is intentionally absent because review mode is `none`. + +## Advisor / Host-Native Record + +- Mode check command: `node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json`. +- Mode result: `B-or-C`; interpreted as Mode B because host `image_gen` is available. +- Garden/API execution expected: no. +- Skill files read: `C:\Users\wangq\.agents\skills\gpt-image-2\SKILL.md`. +- Template/reference files read: `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\concept-scene.md`; `C:\Users\wangq\.agents\skills\gpt-image-2\references\scenes-and-illustrations\minimalist-mood-scene.md`. +- Project source files read: `AGENTS.md`; `VIDEO_WORKBENCH.md`; `docs/workflows/project-lifecycle.md`; `docs/workflows/slide-task-orchestration.md`; `docs/workflows/slide-task-templates.md`; `handoff/2026-06-25-on-the-road-formal-production-handoff.md`; `project.md`; `execution-plan.md`; `slides/slides.md`; task plan/status; `visual-system.md`; `style-prompt-master.md`; `character-reference-registry.md`; S-50 intake row. + +## Final Prompt + +Use Image 1 as the night-road-fire style and mood reference. Use Image 4 as the walking-back character reference: preserve the long dark worn coat, old backpack, shoulder line, back silhouette, boots, and steady walking posture. + +Create a 16:9 photorealistic cinematic documentary road-film still. The traveler has left his small roadside fire and is walking away down the dark wet night road, seen from behind. His old backpack is on his shoulders, his long dark coat and boots are road-worn, and his posture is tired but capable. The small fire remains behind him or side-behind him as a restrained orange glow, not a camp destination. Ahead, the muddy road continues into darkness, with maybe one or two weak far fire points barely visible in the distance. + +Composition: wide road-film frame, road and darkness larger than the person. Traveler about 1/10 to 1/7 of frame, off-center, back view, midground or lower third. Fire about 3% to 6% of frame, behind or beside him, clearly not the endpoint. The road must continue beyond him and pull the eye forward. + +Style: real documentary texture with restrained poetic realism, low saturation, cold gray, earth brown, dark khaki, coal black, wet dark blue-gray. Warmth appears only as small local firelight. Wet mud, puddles, ruts, stones, coat texture, backpack wear, and boot dirt should feel physically real. Emotion: fire is not the destination; it is the reason to continue. + +## Negative / Avoid + +Avoid final arrival, group, reunion, farewell wave, victory pose, heroic centered stance, motivational poster, bright path of destiny, broad sunrise, warm sunset, cozy campsite, tent, camping chairs, large fire, fantasy glow, neon, cyberpunk, UI, readable text, labels, logos, watermarks. + +## Output Intent + +- Save selected host-generated image to `projects/2026-06-23-在路上/slides/s50/s50-v1-image.png`. +- Record as `production-pass / final production direct-child` in `slides/slides.md` after generation succeeds. diff --git a/projects/2026-06-23-在路上/slides/slides.md b/projects/2026-06-23-在路上/slides/slides.md index 5394e69..687883e 100644 --- a/projects/2026-06-23-在路上/slides/slides.md +++ b/projects/2026-06-23-在路上/slides/slides.md @@ -27,6 +27,7 @@ Source: | Unit | Source Shot | Segment / Time | Function | Style Anchor | Batch | Status | Notes | | --- | --- | --- | --- | --- | --- | --- | --- | +| s00 | S-00 | Cover / before 0:00 | Textless MV cover base: muddy road, broken starlight puddle, small back-view traveler, and restrained distant fire with title-safe dark negative space for post-production. | C / D / cover | production-light, design-gated | production-pass | `s00-v1-design.md`, `s00-v1-design-en.md`, `s00-v1-brief.md`, `s00-v1-prompt.md`, and `s00-v1-image.png` generated after user-approved design. No `s00-v1-review.md`; image is a textless base for later manual title work. | | s01 | S-01 | Prelude / 0:00-0:16 | World appears before person; cold frosted muddy road, no person. | A | anchor-test, priority | accepted | `s01-v1-image.png` accepted by user. | | s02 | S-02 | Prelude / 0:00-0:16 | Tiny traveler enters the world along the muddy road. | A | deferred | accepted / R4 complete | `s02-v1-image.png` generated in broad run; keep as historical evidence only. `s02-v2-brief.md`, `s02-v2-prompt.md`, `s02-v2-image.png`, and return exist from `session-05`; no `s02-v2-review.md`. User accepted `s02-v2-image.png`. Mechanism note: brief/packet were written by `Agent S-02` rather than prewritten by the visible child session; visible child session completed host-native generation after worker reference-attachment blocker. | | s03 | S-03 | Prelude / 0:00-0:16 | Old boot and backpack detail connects body to road. | character detail | deferred | accepted / R4 complete | `s03-v1-image.png` generated in broad run; keep as historical evidence only. `s03-v2-brief.md`, `s03-v2-prompt.md`, `s03-v2-image.png`, and return exist from `session-05`; no `s03-v2-review.md`. User accepted `s03-v2-image.png`. Mechanism note: brief/packet were written by `Agent S-03` rather than prewritten by the visible child session; visible child session completed host-native generation after worker reference-attachment blocker. | @@ -52,36 +53,40 @@ Source: | s23 | S-23 | Chorus / 1:59-3:24 | Traveler crosses large muddy ground. | A | deferred | accepted / session-11 retry callback pass | `s23-v1-image.png` generated in broad run; keep as historical evidence only. `s23-v2` is incomplete evidence from paused `session-10`. `session-11` generated append-only `s23-v3-image.png`; completion gate `pass`, no `s23-v3-review.md`, image SHA256 `4A3371B33E6FD555FF4BECAFE18B4208029C2C6F6BFD710B33C74203661FCE69`. User accepted `s23-v3-image.png`. | | s24 | S-24 | Chorus / 1:59-3:24 | Huge unnamed wilderness nearly swallows person. | A | priority | accepted / session-11 retry callback pass | `s24-v1-image.png` generated in broad run; keep as historical evidence only. `session-11` generated `s24-v2-image.png`; completion gate `pass`, no `s24-v2-review.md`, image SHA256 `4E0C8D3192ABE86A4A75CB453D86E74AF2968264CC4F75B5A0E0B4635619ABC1`. User accepted `s24-v2-image.png`. | | s25 | S-25 | Chorus / 1:59-3:24 | Traveler wipes dust from coat/face edge. | character detail | deferred | accepted | `s25-v1-image.png` generated in broad run without prompt/reference trace; keep as candidate evidence only. `s25-v2-prompt.md`, `s25-v2-image.png`, and return exist from `Agent S-25` `019ef803-e757-7041-8f4a-fc9dfce23bbd`; no `s25-v2-review.md`; prompt records Advisor, visible side/front/body refs, and style A adapted for dust/detail. User accepted `s25-v2-image.png`. | -| s26 | S-26 | Chorus / 1:59-3:24 | Night water/starlight begins but is not full reveal. | C | priority | generated | `s26-v1-image.png` generated in broad run; awaiting batch review. | -| s27 | S-27 | Chorus / 1:59-3:24 | Traveler stops before mud/water. | A / C | deferred | generated | `s27-v1-image.png` generated in broad run; awaiting batch review. | -| s28 | S-28 | Chorus / 1:59-3:24 | Bridge/round structure shadow suggests gear-like night. | B | priority | generated | `s28-v1-image.png` generated in broad run; awaiting batch review. | -| s29 | S-29 | Chorus / 1:59-3:24 | Boot restarts one step, mud splashes lightly. | character detail | deferred | generated | `s29-v1-image.png` generated in broad run; awaiting batch review. | -| s30 | S-30 | Chorus / 1:59-3:24 | Thin restrained dawn edge lights road. | A | deferred | generated | `s30-v1-image.png` generated in broad run; awaiting batch review. | -| s31 | S-31 | Chorus / 1:59-3:24 | Traveler reflection in puddle broken by water/wind. | C | deferred | generated | `s31-v1-image.png` generated in broad run; awaiting batch review. | -| s32 | S-32 | Chorus / 1:59-3:24 | Faint second footprint or far fire hints at others. | D / motif | deferred | generated | `s32-v1-image.png` generated in broad run; awaiting batch review. | +| s26 | S-26 | Chorus / 1:59-3:24 | Night water/starlight begins but is not full reveal. | C | priority | production-pass / session-12 callback pass | `s26-v1-image.png` generated in broad run; keep as historical evidence only. `session-12` generated append-only `s26-v2-image.png`; completion gate `pass`, no `s26-v2-review.md`, image SHA256 `EBF7D577932B43A947BC4AF219A1296E842B0BFA59632649D2C21375CBB50F4B`. | +| s27 | S-27 | Chorus / 1:59-3:24 | Traveler stops before mud/water. | A / C | deferred | production-pass / session-12 callback pass | `s27-v1-image.png` generated in broad run; keep as historical evidence only. `session-12` generated append-only `s27-v2-image.png`; completion gate `pass`, no `s27-v2-review.md`, image SHA256 `63CA36874A47109F2D1E9869D6D556C0C0E4E9D85199DF5E05EE1012CC63CB72`. | +| s28 | S-28 | Chorus / 1:59-3:24 | Bridge/round structure shadow suggests gear-like night. | B | priority | production-pass / session-12 callback pass | `s28-v1-image.png` generated in broad run; keep as historical evidence only. `session-12` generated append-only `s28-v2-image.png`; completion gate `pass`, no `s28-v2-review.md`, image SHA256 `F3E262C882D08FCA16B1FC52D6852167F8DA4850C846480F61352EE84C5CCA5E`. | +| s29 | S-29 | Chorus / 1:59-3:24 | Boot restarts one step, mud splashes lightly. | character detail | deferred | production-pass / session-13 direct-child | `s29-v1-image.png` generated in broad run; keep as historical evidence only. `session-13` generated append-only `s29-v2-brief.md`, `s29-v2-prompt.md`, and `s29-v2-image.png` directly in child session without Agent/SubAgent; review mode none. | +| s30 | S-30 | Chorus / 1:59-3:24 | Thin restrained dawn edge lights road. | A | deferred | production-pass / session-13 direct-child | `s30-v1-image.png` generated in broad run; keep as historical evidence only. `session-13` generated append-only `s30-v2-brief.md`, `s30-v2-prompt.md`, and `s30-v2-image.png` directly in child session without Agent/SubAgent; review mode none. | +| s31 | S-31 | Chorus / 1:59-3:24 | Traveler reflection in puddle broken by water/wind. | C | deferred | production-pass / session-13 direct-child | `s31-v1-image.png` generated in broad run; keep as historical evidence only. `session-13` generated append-only `s31-v2-brief.md`, `s31-v2-prompt.md`, and `s31-v2-image.png` directly in child session without Agent/SubAgent; review mode none. | +| s32 | S-32 | Chorus / 1:59-3:24 | Faint second footprint or far fire hints at others. | D / motif | deferred | production-pass / capacity validation accepted | `s32-v1-image.png` generated in broad run; keep as historical evidence. `s32-v2-brief.md`, `s32-v2-prompt.md`, and `s32-v2-image.png` generated in production capacity validation session-01 and accepted by user. | | s33 | S-33 | Chorus / 1:59-3:24 | Short half-side face, jaw tight, still has fire. | character state | deferred | accepted / mechanism pass | `s33-v1-image.png` generated in broad run; keep as candidate evidence only. `s33-v2-brief.md`, `s33-v2-prompt.md`, and `s33-v2-image.png` exist from clean `session-04` with `Agent S-33` `019ef87a-3aa1-75a2-a97c-de8f57e234e1`; no `s33-v2-review.md`. Prompt records `gpt-image-2` Advisor, A/D-adjacent style calibration, and loaded side/front/master character refs. User accepted `s33-v2-image.png`. | -| s34 | S-34 | Chorus / 1:59-3:24 | Worn heel, mud, stones, continued step. | character detail | deferred | generated | `s34-v1-image.png` generated before interruption; bookkeeping repaired, awaiting batch review. | -| s35 | S-35 | Chorus / 1:59-3:24 | Traveler turns toward a darker but clearer road. | A | deferred | not started | Direction forms through action, not revelation. | -| s36 | S-36 | Chorus / 1:59-3:24 | Road enters dark wind/sand; unknown remains. | A | deferred | not started | Do not resolve too early. | -| s37 | S-37 | Chorus / 1:59-3:24 | Road dominates frame, person at edge, faint fire far away. | A / D | deferred | not started | Road becomes protagonist. | -| s38 | S-38 | Chorus / 1:59-3:24 | Small night-road fire with traveler alone. | D | priority | not started | Emotion warms without becoming lively. | -| s39 | S-39 | Guitar solo / 3:24-3:58 | No person, wind over muddy road and puddle. | A / C | deferred | not started | Breath and silence. | -| s40 | S-40 | Guitar solo / 3:24-3:58 | High wide shot; tiny person moves through large world. | A / B | deferred | not started | Documentary observation returns. | -| s41 | S-41 | Guitar solo / 3:24-3:58 | Night puddle with hinted star points. | C | priority | not started | Prepares final C reveal. | -| s42 | S-42 | Guitar solo / 3:24-3:58 | Far fire and near weak star puddle, person passes between. | C / D | deferred | not started | Connects star and fire. | -| s43 | S-43 | Guitar solo / 3:24-3:58 | Traveler stops before puddle; near still frame. | C | deferred | not started | Pause before core line. | -| s44 | S-44 | Final four lines / 3:58-4:16 | Traveler keeps walking, no looking back. | A / D | deferred | not started | Action replaces slogan. | -| s45 | S-45 | Final four lines / 3:58-4:16 | Back view passes old footprints without looking back. | A / motif | deferred | not started | Leads into C reveal. | +| s34 | S-34 | Chorus / 1:59-3:24 | Worn heel, mud, stones, continued step. | character detail | deferred | production-pass / capacity validation accepted | `s34-v1-image.png` generated before interruption; keep as historical evidence. `s34-v2-brief.md`, `s34-v2-prompt.md`, and `s34-v2-image.png` generated in production capacity validation session-01 and accepted by user. | +| s35 | S-35 | Chorus / 1:59-3:24 | Traveler turns toward a darker but clearer road. | A | deferred | production-pass / capacity validation accepted | `s35-v1-brief.md`, `s35-v1-prompt.md`, and `s35-v1-image.png` generated in production capacity validation session-01 and accepted by user. Direction forms through action, not revelation. | +| s36 | S-36 | Chorus / 1:59-3:24 | Road enters dark wind/sand; unknown remains. | A | deferred | production-pass / capacity validation accepted | `s36-v1-brief.md`, `s36-v1-prompt.md`, and `s36-v1-image.png` generated in production capacity validation session-01 and accepted by user. Do not resolve too early. | +| s37 | S-37 | Chorus / 1:59-3:24 | Road dominates frame, person at edge, faint fire far away. | A / D | deferred | production-pass / capacity validation accepted | `s37-v1-brief.md`, `s37-v1-prompt.md`, and `s37-v1-image.png` generated in production capacity validation session-01 and accepted by user. Road becomes protagonist. | +| s38 | S-38 | Chorus / 1:59-3:24 | Small night-road fire with traveler alone. | D | priority | production-pass / 8-slide validation accepted | `s38-v1-brief.md`, `s38-v1-prompt.md`, and `s38-v1-image.png` generated in 8-slide compression validation session-01 and accepted into production-pass by production-stage default. | +| s39 | S-39 | Guitar solo / 3:24-3:58 | No person, wind over muddy road and puddle. | A / C | deferred | production-pass / 8-slide validation accepted | `s39-v1-brief.md`, `s39-v1-prompt.md`, and `s39-v1-image.png` generated in 8-slide compression validation session-01 and accepted into production-pass by production-stage default. | +| s40 | S-40 | Guitar solo / 3:24-3:58 | High wide shot; tiny person moves through large world. | A / B | deferred | production-pass / 8-slide validation accepted | `s40-v1-brief.md`, `s40-v1-prompt.md`, and `s40-v1-image.png` generated in 8-slide compression validation session-01 and accepted into production-pass by production-stage default. | +| s41 | S-41 | Guitar solo / 3:24-3:58 | Night puddle with hinted star points. | C | priority | production-pass / 8-slide validation accepted | `s41-v1-brief.md`, `s41-v1-prompt.md`, and `s41-v1-image.png` generated in 8-slide compression validation session-01 and accepted into production-pass by production-stage default. | +| s42 | S-42 | Guitar solo / 3:24-3:58 | Far fire and near weak star puddle, person passes between. | C / D | deferred | production-pass / 8-slide validation accepted | `s42-v1-brief.md`, `s42-v1-prompt.md`, and `s42-v1-image.png` generated in 8-slide compression validation session-01 and accepted into production-pass by production-stage default. | +| s43 | S-43 | Guitar solo / 3:24-3:58 | Traveler stops before puddle; near still frame. | C | deferred | production-pass / 8-slide validation accepted | `s43-v1-brief.md`, `s43-v1-prompt.md`, and `s43-v1-image.png` generated in 8-slide compression validation session-01 and accepted into production-pass by production-stage default. | +| s44 | S-44 | Final four lines / 3:58-4:16 | Traveler keeps walking, no looking back. | A / D | deferred | production-pass / 8-slide validation accepted | `s44-v1-brief.md`, `s44-v1-prompt.md`, and `s44-v1-image.png` generated in 8-slide compression validation session-01 and accepted into production-pass by production-stage default. | +| s45 | S-45 | Final four lines / 3:58-4:16 | Back view passes old footprints without looking back. | A / motif | deferred | production-pass / 8-slide validation accepted | `s45-v1-brief.md`, `s45-v1-prompt.md`, and `s45-v1-image.png` generated in 8-slide compression validation session-01 and accepted into production-pass by production-stage default. | | s46 | S-46 | Final four lines / 3:58-4:16 | Muddy puddle with broken star reflection and boot edge. | C | anchor-test, expanded, priority | accepted | `s46-v1-image.png` accepted by user. Soul shot; mud and stars both true. | | s47 | S-47 | Outro / 4:16-4:54 | Transition from muddy starlight to night road fire. | D | anchor-test, priority | accepted | `s47-v1-image.png` accepted by user. Restrained D warmth after C. | -| s48 | S-48 | Outro / 4:16-4:54 | Second/third weak fire points appear in darkness. | D | priority | not started | Others become visible only as fire. | -| s49 | S-49 | Outro / 4:16-4:54 | Traveler quietly looks toward distant fire, does not approach. | D | deferred | not started | Confirmation without reunion. | -| s50 | S-50 | Outro / 4:16-4:54 | Traveler leaves his fire and continues. | D | deferred | not started | Fire is not destination. | +| s48 | S-48 | Outro / 4:16-4:54 | Second/third weak fire points appear in darkness. | D | priority | production-pass / final production direct-child | `s48-v1-brief.md`, `s48-v1-prompt.md`, and `s48-v1-image.png` generated in final production S48-S50 direct child; review mode none. Others become visible only as fire. | +| s49 | S-49 | Outro / 4:16-4:54 | Traveler quietly looks toward distant fire, does not approach. | D | deferred | production-pass / final production direct-child | `s49-v1-brief.md`, `s49-v1-prompt.md`, and `s49-v1-image.png` generated in final production S48-S50 direct child; review mode none. Confirmation without reunion. | +| s50 | S-50 | Outro / 4:16-4:54 | Traveler leaves his fire and continues. | D | deferred | production-pass / final production direct-child | `s50-v1-brief.md`, `s50-v1-prompt.md`, and `s50-v1-image.png` generated in final production S48-S50 direct child; review mode none. Fire is not destination. | | s51 | S-51 | Outro / 4:16-4:54 | Final wide: night road, several far fires, tiny traveler. | D | anchor-test, expanded, priority | accepted | `s51-v1-image.png` accepted by user. Ending remains unfinished and restrained. | ## Current Active Batch -R6 Phase 2 retry outputs for `S-20/S-22/S-23/S-24` are accepted. Broad production is waiting for the next user production decision. +Cover output `S-00` is accepted into production-pass as a design-gated textless MV cover base. R7 production-light outputs for `S-26/S-27/S-28/S-29/S-30/S-31` are accepted. Capacity validation outputs for `S-32/S-34/S-35/S-36/S-37`, 8-slide validation outputs for `S-38/S-39/S-40/S-41/S-42/S-43/S-44/S-45`, and final production direct-child outputs for `S-48/S-49/S-50` are accepted into production-pass. There is no remaining not-started shot-image production scope in the current 51-shot table. + +Accepted cover output: + +- `s00/s00-v1-image.png` (production-pass; design-gated textless MV cover base; no review file) Accepted anchor / trial shots: @@ -119,12 +124,49 @@ Accepted R6 callback-only outputs: - `s23/s23-v3-image.png` (accepted; Phase 2 retry callback package) - `s24/s24-v2-image.png` (accepted; Phase 2 retry callback package) -Latest clean mechanism pass: +Accepted R7 production-light outputs: -- Task: `tasks/2026-06-24-character-consistency/session-06/` -- Child session: `019ef907-16b7-7a43-9924-227892aa62d3` -- Internal slide workers: `Agent S-07` `019ef90a-422b-7333-9021-f19b8dd1c0b7`; retry `Agent S-08` `019ef927-d480-7252-9650-8b0670f877d7`; `Agent S-09` `019ef92d-fa13-7053-aec9-fa8abb75f8d2` -- Review files: intentionally absent for all three accepted outputs. +- `s26/s26-v2-image.png` (accepted; `session-12` production-light package) +- `s27/s27-v2-image.png` (accepted; `session-12` production-light package) +- `s28/s28-v2-image.png` (accepted; `session-12` production-light package) +- `s29/s29-v2-image.png` (accepted; `session-13` direct-child package) +- `s30/s30-v2-image.png` (accepted; `session-13` direct-child package) +- `s31/s31-v2-image.png` (accepted; `session-13` direct-child package) + +Accepted production capacity validation outputs: + +- `s32/s32-v2-image.png` (accepted; 5-slide direct-child validation package) +- `s34/s34-v2-image.png` (accepted; 5-slide direct-child validation package) +- `s35/s35-v1-image.png` (accepted; 5-slide direct-child validation package) +- `s36/s36-v1-image.png` (accepted; 5-slide direct-child validation package) +- `s37/s37-v1-image.png` (accepted; 5-slide direct-child validation package) + +Accepted 8-slide compression validation outputs: + +- `s38/s38-v1-image.png` (production-pass; 8-slide direct-child validation package) +- `s39/s39-v1-image.png` (production-pass; 8-slide direct-child validation package) +- `s40/s40-v1-image.png` (production-pass; 8-slide direct-child validation package) +- `s41/s41-v1-image.png` (production-pass; 8-slide direct-child validation package) +- `s42/s42-v1-image.png` (production-pass; 8-slide direct-child validation package) +- `s43/s43-v1-image.png` (production-pass; 8-slide direct-child validation package) +- `s44/s44-v1-image.png` (production-pass; 8-slide direct-child validation package) +- `s45/s45-v1-image.png` (production-pass; 8-slide direct-child validation package) + +Accepted final production outputs: + +- `s48/s48-v1-image.png` (production-pass; final production direct-child package) +- `s49/s49-v1-image.png` (production-pass; final production direct-child package) +- `s50/s50-v1-image.png` (production-pass; final production direct-child package) + +Latest accepted production route: + +- Task: `tasks/2026-06-25-production-capacity-validation/session-01/` +- Child session: `019efc24-96a4-7582-bb05-914b82a726d5` +- Topology: direct child session, no Agent/SubAgent. +- Review mode: none; review files intentionally absent. +- Process note: child writes each slide's brief and prompt first, then generates images in sequence so the package story context stays coherent. +- User-observed performance: callback at about 152K tokens / 58% context; final observed usage about 164K tokens / 63% context for 5 slides. +- Mechanism note: this validation exposed a callback-order issue because the child wrote/checks after callback. Workflow docs now require callback to be the child's final action. - Production route: `gpt-image-2` Advisor prompt/spec plus active style and character references loaded before host image generation. R4 formal-run result: diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/child-session-handoff.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/child-session-handoff.md new file mode 100644 index 0000000..012cb1b --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/child-session-handoff.md @@ -0,0 +1,18 @@ +# Child Session 12 Handoff + +- Status: complete +- Completion gate: pass; append-only versions resolved as `s26-v2`, `s27-v2`, `s28-v2`; no overwrite; expected briefs, packets, prompts, images, and return receipts exist; review files are absent by `review: none`; one terminal worker owned each slide/version; required visible-context references were loaded before host generation. +- Callback: sent to `019efb9f-ece2-78d2-ba6a-4ca1dcfe3244` +- Review mode: none +- Resolved versions: `S-26/s26-v2`, `S-27/s27-v2`, `S-28/s28-v2` +- Image paths: + - `projects/2026-06-23-在路上/slides/s26/s26-v2-image.png` SHA256 `EBF7D577932B43A947BC4AF219A1296E842B0BFA59632649D2C21375CBB50F4B` + - `projects/2026-06-23-在路上/slides/s27/s27-v2-image.png` SHA256 `63CA36874A47109F2D1E9869D6D556C0C0E4E9D85199DF5E05EE1012CC63CB72` + - `projects/2026-06-23-在路上/slides/s28/s28-v2-image.png` SHA256 `F3E262C882D08FCA16B1FC52D6852167F8DA4850C846480F61352EE84C5CCA5E` +- Return receipts: + - `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s26-v2-slide-agent-return.md` + - `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s27-v2-slide-agent-return.md` + - `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s28-v2-slide-agent-return.md` +- Problems: none +- Needs user: none +- Parent summary: R7 production-light `S-26/S-27/S-28` completed through `gpt-image-2` Advisor / host-native generation with visible-context references loaded, no MiniMax/external review, no review files, and compact receipts only. Generated images are production-pass outputs by task rule; user may request an append-only retry later if any result is unusable. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/child-session-plan.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/child-session-plan.md new file mode 100644 index 0000000..a871f62 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/child-session-plan.md @@ -0,0 +1,17 @@ +# Child Session 12 Plan + +- Line: production-light +- Parent task: `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/task-plan.md` +- Parent callback target: `019efb9f-ece2-78d2-ba6a-4ca1dcfe3244` +- Assigned slides: `S-26/s26-v2`, `S-27/s27-v2`, `S-28/s28-v2` +- Running mode: prompt and generate +- Review mode: none +- Write policy: append-only; overwrite allowed: no +- Resolved versions: `s26-v2`, `s27-v2`, `s28-v2`; resolved by scanning `slides/s26/`, `slides/s27/`, and `slides/s28/`, where only `v1` broad-run images existed before this session. +- Shared context: `AGENTS.md`, `VIDEO_WORKBENCH.md`, `docs/workflows/slide-task-orchestration.md`, `docs/workflows/slide-task-templates.md`, `project.md`, `execution-plan.md`, `slides/slides.md`, task plan, style master, character registry, `visual-system/visual-system.md` +- Visual-system files: `visual-system/style/style-prompt-master.md`; `visual-system/characters/character-reference-registry.md` +- Skill mode: `gpt-image-2` check returned `mode=B-or-C`, `recommendation=host-or-advisor`, `garden_mode_enabled=false`, `has_api_key=false` +- Skill references read: `gpt-image-2/SKILL.md`, `references/prompt-writing.md`, `references/scenes-and-illustrations/concept-scene.md`, `references/scenes-and-illustrations/minimalist-mood-scene.md` +- Packet paths: `session-12/packets/s26-v2-slide-agent-packet.md`, `session-12/packets/s27-v2-slide-agent-packet.md`, `session-12/packets/s28-v2-slide-agent-packet.md` +- Stop conditions: any target collision; inability to load required visible-context reference; host image generation unavailable; missing prompt/image/return; duplicate worker; review file produced despite `review: none` + diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s26-v2-slide-agent-packet.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s26-v2-slide-agent-packet.md new file mode 100644 index 0000000..e754439 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s26-v2-slide-agent-packet.md @@ -0,0 +1,87 @@ +# S-26 Slide Agent Packet V2 + +## Status + +- Status: issued by session-12 child session. +- Worker ownership: exactly one terminal worker for `s26-v2`. + +## Target Slide + +- Slide: `S-26` +- Version: `s26-v2` +- Brief: `projects/2026-06-23-在路上/slides/s26/s26-v2-brief.md` + +## Allowed Actions + +- Read the packet, brief, required workflow docs, style master, and character registry. +- Write `projects/2026-06-23-在路上/slides/s26/s26-v2-prompt.md`. +- Generate one image if host image generation is available. +- Write `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s26-v2-slide-agent-return.md`. + +## Input Files + +- `AGENTS.md` +- `docs/workflows/slide-task-orchestration.md` +- `docs/workflows/slide-task-templates.md` +- `projects/2026-06-23-在路上/slides/s26/s26-v2-brief.md` +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + +## Output Files + +- `projects/2026-06-23-在路上/slides/s26/s26-v2-prompt.md` +- `projects/2026-06-23-在路上/slides/s26/s26-v2-image.png` +- `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s26-v2-slide-agent-return.md` + +## Artifact Version / Write Policy + +- Append-only version: `s26-v2` +- Overwrite allowed: no +- Stop on any target collision. + +## Reference Roles + +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png`; C muddy-water / broken-starlight style reference. +- Image 2: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; overall style calibration. +- Character reference: none required. + +## Visual-System Requirements + +- Use style module C. +- Keep C as a beginning / foreshadowing shot, not the full S-46 reveal. + +## Gpt-Image-2 Advisor Rule + +- Use `gpt-image-2` as Advisor / host-native prompt workflow. +- Record the mode check result and template/reference files used. + +## Host Generation Evidence Rule + +- Load required visible-context references or block before generation. +- Record loaded reference roles and final project image path in the return. + +## Generation Rule + +- Running mode: prompt and generate. +- Garden/API expected: no. + +## Review Rule + +- Review mode: none. +- Do not write `s26-v2-review.md`. + +## Completion Gate Rule + +- Expected prompt, image, and return exist. +- No overwrite. +- Review absent. +- Worker lifecycle terminal. + +## Lifecycle Rule + +- This worker handles only `s26-v2` and becomes terminal after the return. + +## Return File + +- `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s26-v2-slide-agent-return.md` + diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s27-v2-slide-agent-packet.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s27-v2-slide-agent-packet.md new file mode 100644 index 0000000..3de59d6 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s27-v2-slide-agent-packet.md @@ -0,0 +1,88 @@ +# S-27 Slide Agent Packet V2 + +## Status + +- Status: issued by session-12 child session. +- Worker ownership: exactly one terminal worker for `s27-v2`. + +## Target Slide + +- Slide: `S-27` +- Version: `s27-v2` +- Brief: `projects/2026-06-23-在路上/slides/s27/s27-v2-brief.md` + +## Allowed Actions + +- Read the packet, brief, required workflow docs, style master, and character registry. +- Write `projects/2026-06-23-在路上/slides/s27/s27-v2-prompt.md`. +- Generate one image if host image generation is available. +- Write `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s27-v2-slide-agent-return.md`. + +## Input Files + +- `AGENTS.md` +- `docs/workflows/slide-task-orchestration.md` +- `docs/workflows/slide-task-templates.md` +- `projects/2026-06-23-在路上/slides/s27/s27-v2-brief.md` +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + +## Output Files + +- `projects/2026-06-23-在路上/slides/s27/s27-v2-prompt.md` +- `projects/2026-06-23-在路上/slides/s27/s27-v2-image.png` +- `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s27-v2-slide-agent-return.md` + +## Artifact Version / Write Policy + +- Append-only version: `s27-v2` +- Overwrite allowed: no +- Stop on any target collision. + +## Reference Roles + +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; overall style calibration. +- Image 2: `projects/2026-06-23-在路上/intake/imgs/A.风景主导型蓝图.png`; landscape-led road composition. +- Image 3: `projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png`; muddy-water / faint starlight cue. +- Image 4: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`; back-view traveler silhouette continuity. + +## Visual-System Requirements + +- Use mixed A / C module. +- Character reference is required because the traveler is visible. + +## Gpt-Image-2 Advisor Rule + +- Use `gpt-image-2` as Advisor / host-native prompt workflow. +- Record the mode check result and template/reference files used. + +## Host Generation Evidence Rule + +- Load required visible-context references or block before generation. +- Record loaded reference roles and final project image path in the return. + +## Generation Rule + +- Running mode: prompt and generate. +- Garden/API expected: no. + +## Review Rule + +- Review mode: none. +- Do not write `s27-v2-review.md`. + +## Completion Gate Rule + +- Expected prompt, image, and return exist. +- No overwrite. +- Review absent. +- Worker lifecycle terminal. + +## Lifecycle Rule + +- This worker handles only `s27-v2` and becomes terminal after the return. + +## Return File + +- `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s27-v2-slide-agent-return.md` + diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s28-v2-slide-agent-packet.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s28-v2-slide-agent-packet.md new file mode 100644 index 0000000..3cfa3b7 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/packets/s28-v2-slide-agent-packet.md @@ -0,0 +1,88 @@ +# S-28 Slide Agent Packet V2 + +## Status + +- Status: issued by session-12 child session. +- Worker ownership: exactly one terminal worker for `s28-v2`. + +## Target Slide + +- Slide: `S-28` +- Version: `s28-v2` +- Brief: `projects/2026-06-23-在路上/slides/s28/s28-v2-brief.md` + +## Allowed Actions + +- Read the packet, brief, required workflow docs, style master, and character registry. +- Write `projects/2026-06-23-在路上/slides/s28/s28-v2-prompt.md`. +- Generate one image if host image generation is available. +- Write `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s28-v2-slide-agent-return.md`. + +## Input Files + +- `AGENTS.md` +- `docs/workflows/slide-task-orchestration.md` +- `docs/workflows/slide-task-templates.md` +- `projects/2026-06-23-在路上/slides/s28/s28-v2-brief.md` +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` + +## Output Files + +- `projects/2026-06-23-在路上/slides/s28/s28-v2-prompt.md` +- `projects/2026-06-23-在路上/slides/s28/s28-v2-image.png` +- `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s28-v2-slide-agent-return.md` + +## Artifact Version / Write Policy + +- Append-only version: `s28-v2` +- Overwrite allowed: no +- Stop on any target collision. + +## Reference Roles + +- Image 1: `projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png`; overall style calibration. +- Image 2: `projects/2026-06-23-在路上/intake/imgs/B.高位观察型蓝图.png`; high-angle city-edge structure pressure. +- Image 3: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png`; walking side silhouette continuity. +- Image 4: `projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png`; walking-away silhouette continuity. + +## Visual-System Requirements + +- Use style module B. +- Character references are required because the traveler is visible. + +## Gpt-Image-2 Advisor Rule + +- Use `gpt-image-2` as Advisor / host-native prompt workflow. +- Record the mode check result and template/reference files used. + +## Host Generation Evidence Rule + +- Load required visible-context references or block before generation. +- Record loaded reference roles and final project image path in the return. + +## Generation Rule + +- Running mode: prompt and generate. +- Garden/API expected: no. + +## Review Rule + +- Review mode: none. +- Do not write `s28-v2-review.md`. + +## Completion Gate Rule + +- Expected prompt, image, and return exist. +- No overwrite. +- Review absent. +- Worker lifecycle terminal. + +## Lifecycle Rule + +- This worker handles only `s28-v2` and becomes terminal after the return. + +## Return File + +- `projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s28-v2-slide-agent-return.md` + diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s26-v2-slide-agent-return.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s26-v2-slide-agent-return.md new file mode 100644 index 0000000..9172711 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s26-v2-slide-agent-return.md @@ -0,0 +1,16 @@ +# S-26 Slide Agent Return V2 + +- Status: complete +- Slide/version: `S-26` / `s26-v2` +- Prompt: `projects/2026-06-23-在路上/slides/s26/s26-v2-prompt.md` +- Image: `projects/2026-06-23-在路上/slides/s26/s26-v2-image.png` +- Review mode: none +- Completeness gate: pass. Prompt and image files exist; `s26-v2-review.md` is absent as instructed; append-only `s26-v2` targets were checked before writing; no overwrite was performed. +- Advisor / host-native: `gpt-image-2` mode check returned `B-or-C`; interpreted as Mode B because host image generation was available. Garden/API expected: no. Garden/API used: no. +- Reference roles loaded: Image 1 loaded as C repaired muddy-water / broken-starlight style reference; Image 2 loaded as overall low-saturation documentary road-film calibration board. Character reference: none required. +- Host generated-image source: `C:\Users\wangq\.codex\generated_images\019efbb8-efe6-7741-be0a-64e4188ac751\ig_0faf3d58c54f27c8016a3c58e83d70819191492b8dff5ef7cb.png` +- Project image SHA256: `EBF7D577932B43A947BC4AF219A1296E842B0BFA59632649D2C21375CBB50F4B` +- Local visual check: image is a 16:9 low ground-level night mud / puddle frame with dirty water, grit, stones, faint broken cold starlight, and no visible full character. +- Agent lifecycle: terminal / closed after this return. +- Blockers: none. +- Parent summary: `s26-v2` prompt and one host-generated image were written using Image 1 and Image 2 visible-context references; review intentionally suppressed; completion gate passed. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s27-v2-slide-agent-return.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s27-v2-slide-agent-return.md new file mode 100644 index 0000000..4ea5cb4 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s27-v2-slide-agent-return.md @@ -0,0 +1,15 @@ +# S-27 Slide Agent Return V2 + +- Status: complete +- Slide/version: `s27-v2` +- Prompt: `projects/2026-06-23-在路上/slides/s27/s27-v2-prompt.md` +- Image: `projects/2026-06-23-在路上/slides/s27/s27-v2-image.png` +- Review mode: none; `projects/2026-06-23-在路上/slides/s27/s27-v2-review.md` intentionally absent +- Completeness gate: pass; append-only `s27-v2`, no overwrite, prompt exists, image exists, review absent as instructed, return recorded +- Advisor / host-native: `gpt-image-2` mode check returned `B-or-C`; Garden/API expected: no; host image generation used once +- Reference roles loaded: Image 1 style calibration board; Image 2 landscape-led road composition; Image 3 muddy water / faint starlight cue; Image 4 back-view traveler identity/silhouette +- Host generated-image source: `C:\Users\wangq\.codex\generated_images\019efbbe-635a-7761-80f1-56eee5effd07\ig_022de87cade04ddc016a3c5a46f7c88193a1d1b104f20dc1f1.png` +- Project image evidence: `1672x941`, SHA256 `63CA36874A47109F2D1E9869D6D556C0C0E4E9D85199DF5E05EE1012CC63CB72` +- Agent lifecycle: terminal / closed after this return +- Blockers: none +- Parent summary: `s27-v2` completed in production-light mode with no review file; prompt and image are written, references Images 1-4 were used through the host-native visible-context route, and the worker is terminal. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s28-v2-slide-agent-return.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s28-v2-slide-agent-return.md new file mode 100644 index 0000000..3950968 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-12/returns/s28-v2-slide-agent-return.md @@ -0,0 +1,16 @@ +# S-28 Slide Agent Return V2 + +- Status: complete. +- Slide/version: `S-28 / s28-v2`. +- Prompt: `projects/2026-06-23-在路上/slides/s28/s28-v2-prompt.md`. +- Image: `projects/2026-06-23-在路上/slides/s28/s28-v2-image.png`. +- Review mode: none; `projects/2026-06-23-在路上/slides/s28/s28-v2-review.md` intentionally absent. +- Completeness gate: pass. Prompt and image exist; return exists after this write; append-only `s28-v2` targets were missing before write; no overwrite performed; review absent as instructed. +- Advisor / host-native: `gpt-image-2` mode check returned `B-or-C`, interpreted as host-native Mode B because host image generation was available. Garden/API expected: no; Garden/API not used. +- Reference roles loaded: Image 1 style anchor board / overall calibration; Image 2 high-angle city-edge pressure; Image 3 walking side silhouette continuity; Image 4 walking-away silhouette and backpack/coat continuity. +- Host generated-image source: `C:\Users\wangq\.codex\generated_images\019efbc4-9000-7580-a2b7-7d67a7e8cd19\ig_0acd0868576f443a016a3c5bc228a48191a5fc31ca9779c1a7.png`. +- Project image check: PNG readable, 1672x941, SHA256 `F3E262C882D08FCA16B1FC52D6852167F8DA4850C846480F61352EE84C5CCA5E`. +- Local visual note: massive circular bridge/overpass shadow dominates the wet city-edge ground; traveler remains small and secondary. +- Agent lifecycle: terminal / closed after this return. +- Blockers: none. +- Parent summary: `s28-v2` prompt and image were produced through the Advisor / host-native path with Image 1-4 visible references; review was intentionally suppressed; completion gate passed. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-13/child-session-handoff.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-13/child-session-handoff.md new file mode 100644 index 0000000..1a43c5a --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-13/child-session-handoff.md @@ -0,0 +1,14 @@ +# Session 13 Handoff + +- status: complete +- topology: direct-child, no Agent/SubAgent +- resolved versions: `s29-v2`, `s30-v2`, `s31-v2` +- image paths: + - `projects/2026-06-23-在路上/slides/s29/s29-v2-image.png` + - `projects/2026-06-23-在路上/slides/s30/s30-v2-image.png` + - `projects/2026-06-23-在路上/slides/s31/s31-v2-image.png` +- review mode: none +- problems: none recorded +- needs_user: none +- observed elapsed time: not precisely exposed by runtime +- observed context/token usage: not visible to this child session diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/task-log.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/task-log.md index c0ddaab..09d4d47 100644 --- a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/task-log.md +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/task-log.md @@ -2,9 +2,11 @@ ## Status -- Current status: R6 Phase 2 retry callback completed and user accepted `S-20/S-22/S-23/S-24` +- Current status: R7 production-light direct-child test accepted; `session-13` no-Agent package for `S-29/S-30/S-31` is the formal production default - Running mode: `prompt and generate` -- Authorized current slides: none; wait for new-session R6 start +- Authorized current slides: none active; next production scope pending user decision +- Active line: `production-light` +- Review mode: `none` - Deprecated child session id: `019ef77d-552b-7343-abe4-88921b35d7ea` (do not reuse) - Session-02 child-session id: `019ef7ca-413a-7662-9633-7fb126b7ef35` - Session-02 resolved version: `s21-v2` (runtime fallback; no prompt/image) @@ -48,8 +50,19 @@ - Session-11 dispatched scope: `S-20`, `S-22`, `S-23`, `S-24` - Session-11 dispatch mode: callback-only; parent did not poll the child thread - Session-11 result: completion gate `pass`; generated `s20-v3`, `s22-v3`, `s23-v3`, `s24-v2`; user accepted all four images +- Session-12 result: generated `s26-v2`, `s27-v2`, `s28-v2`; user accepted all three images; user observed 25m48s / 134K tokens +- Session-13 result: direct child session without Agent/SubAgent generated `s29-v2`, `s30-v2`, `s31-v2`; user accepted all three images; user observed 10m32s / 114K tokens - Write policy: append-only next slide version; overwrite disabled +## R7 Production-Light Dispatch / Callback Receipts + +- 2026-06-25 06:13 dispatch: `session-12` thread `019efbb4-b366-7de3-8e02-01183b3d0b13`, scope `S-26/S-27/S-28`, line `production-light`, review `none`, expected versions `s26-v2/s27-v2/s28-v2`, callback target `019efb9f-ece2-78d2-ba6a-4ca1dcfe3244`, overwrite allowed `no`. +- 2026-06-25 06:42 callback: status `complete`, completion gate `pass`, resolved versions `s26-v2/s27-v2/s28-v2`, image paths `slides/s26/s26-v2-image.png`, `slides/s27/s27-v2-image.png`, `slides/s28/s28-v2-image.png`, handoff `session-12/child-session-handoff.md`, needs user `none`. +- 2026-06-25 user metric: `session-12` took 25m48s and 134K tokens for 3 slides; images accepted. +- 2026-06-25 06:55 dispatch: `session-13` thread `019efbd8-f75d-7512-9952-aa3528fde581`, scope `S-29/S-30/S-31`, line `production-light`, topology `direct-child/no-Agent`, review `none`, expected versions `s29-v2/s30-v2/s31-v2`, callback target `019efb9f-ece2-78d2-ba6a-4ca1dcfe3244`, overwrite allowed `no`. +- 2026-06-25 07:07 callback: `session-13`, status `complete`, topology `direct-child/no-Agent`, resolved versions `s29-v2/s30-v2/s31-v2`, image paths `slides/s29/s29-v2-image.png`, `slides/s30/s30-v2-image.png`, `slides/s31/s31-v2-image.png`, handoff `session-13/child-session-handoff.md`, problems `none`, needs user `none`. +- 2026-06-25 user metric: `session-13` took 10m32s and 114K tokens for 3 slides; images accepted. Direct-child/no-Agent is accepted as the formal production default. + ## Timeline - 2026-06-24: Task initialized from handoff and user instruction. @@ -128,6 +141,8 @@ - 2026-06-24: Parent local post-processing verified `session-11` handoff, prompt/image/return files for `s20-v3`, `s22-v3`, `s23-v3`, and `s24-v2`, and absence of all four review files. Image SHA256 values: `s20-v3` `062BA727BA2D98CEF9EF5FDD142E2378C03F52B369C6923EFD9007AD3CD5A99B`; `s22-v3` `0109FF60DFC43CE4CB834981C17AC55B0986F73DFB644B51331762A62F8333FD`; `s23-v3` `4A3371B33E6FD555FF4BECAFE18B4208029C2C6F6BFD710B33C74203661FCE69`; `s24-v2` `4E0C8D3192ABE86A4A75CB453D86E74AF2968264CC4F75B5A0E0B4635619ABC1`. - 2026-06-24: Parent local evidence confirmed MiniMax MCP, external vision/review services, CCPE, GPT V2, and web search were not called in `session-11` returns/ledger. - 2026-06-25: User visually accepted `s20-v3-image.png`, `s22-v3-image.png`, `s23-v3-image.png`, and `s24-v2-image.png`. R6 Phase 2 retry is accepted for both callback-only mechanism and image results. +- 2026-06-25: R7 production-light package `session-12` completed `s26-v2`, `s27-v2`, and `s28-v2`; user accepted all three images and observed 25m48s / 134K tokens. +- 2026-06-25: R7 direct-child package `session-13` completed `s29-v2`, `s30-v2`, and `s31-v2` without Agent/SubAgent, packet/return receipt, review, or final verification pass; user accepted all three images and observed 10m32s / 114K tokens. This becomes the formal production route. ## Child Session Results @@ -163,6 +178,8 @@ - `session-09` result: callback woke parent thread `019ef9b5-cc39-7a42-91b1-f9b0852dc95e`; completion gate `pass`; `s16-v3` prompt/image/return exist; no review file; user accepted the image. - `session-10` child thread: `019ef9d0-4914-7ea0-8935-bfd9bfc1aad4`; R6 Phase 2 first 4-shot package for `S-19/S-20/S-22/S-23`; user paused after `S-20` worker boundary violation. `s19-v2` accepted; `s20-v2`, `s22-v2`, and `s23-v2` are incomplete evidence only. - `session-11` child thread: `019ef9f8-8c34-77a3-82af-9a25beb2c723`; R6 Phase 2 retry 4-shot package for `S-20/S-22/S-23/S-24`; callback received, completion gate `pass`, generated `s20-v3`, `s22-v3`, `s23-v3`, `s24-v2`; user accepted all four images. +- 2026-06-25 06:40 callback: `session-12`, `S-26/S-27/S-28`, production-light, review `none`, completion gate `pass`, generated `s26-v2`, `s27-v2`, `s28-v2`; image paths recorded in `session-12/child-session-handoff.md`; needs_user: none. +- `session-13` child thread: `019efbd8-f75d-7512-9952-aa3528fde581`; R7 direct-child production-light package for `S-29/S-30/S-31`; no Agent/SubAgent; generated `s29-v2`, `s30-v2`, `s31-v2`; user accepted all three images. ## Slide Status Summary @@ -187,6 +204,12 @@ | `s22` | accepted / session-11 retry callback pass | `s22-v2` remains incomplete evidence from paused `session-10`. `session-11` generated append-only `s22-v3` with prompt/image/return, no review file, and no external vision/MiniMax use. Image SHA256 `0109FF60DFC43CE4CB834981C17AC55B0986F73DFB644B51331762A62F8333FD`. User accepted `s22-v3-image.png`. | | `s23` | accepted / session-11 retry callback pass | `s23-v2` remains incomplete evidence from paused `session-10`. `session-11` generated append-only `s23-v3` with prompt/image/return, no review file, and no external vision/MiniMax use. Image SHA256 `4A3371B33E6FD555FF4BECAFE18B4208029C2C6F6BFD710B33C74203661FCE69`. User accepted `s23-v3-image.png`. | | `s24` | accepted / session-11 retry callback pass | `s24-v1-image.png` is broad-run historical evidence. `session-11` generated `s24-v2` with prompt/image/return, no review file, and no external vision/MiniMax use. Image SHA256 `4E0C8D3192ABE86A4A75CB453D86E74AF2968264CC4F75B5A0E0B4635619ABC1`. User accepted `s24-v2-image.png`. | +| `s26` | production-pass / session-12 callback pass | `s26-v1-image.png` is broad-run historical evidence. `session-12` generated `s26-v2` with prompt/image/return, no review file, and no external vision/MiniMax use. Image SHA256 `EBF7D577932B43A947BC4AF219A1296E842B0BFA59632649D2C21375CBB50F4B`. | +| `s27` | production-pass / session-12 callback pass | `s27-v1-image.png` is broad-run historical evidence. `session-12` generated `s27-v2` with prompt/image/return, no review file, and no external vision/MiniMax use. Image SHA256 `63CA36874A47109F2D1E9869D6D556C0C0E4E9D85199DF5E05EE1012CC63CB72`. | +| `s28` | production-pass / session-12 callback pass | `s28-v1-image.png` is broad-run historical evidence. `session-12` generated `s28-v2` with prompt/image/return, no review file, and no external vision/MiniMax use. Image SHA256 `F3E262C882D08FCA16B1FC52D6852167F8DA4850C846480F61352EE84C5CCA5E`. | +| `s29` | production-pass / session-13 direct-child | `s29-v1-image.png` is broad-run historical evidence. `session-13` generated `s29-v2` directly in the child session without Agent/SubAgent; brief, prompt, and image exist; review mode `none`. User accepted the image. | +| `s30` | production-pass / session-13 direct-child | `s30-v1-image.png` is broad-run historical evidence. `session-13` generated `s30-v2` directly in the child session without Agent/SubAgent; brief, prompt, and image exist; review mode `none`. User accepted the image. | +| `s31` | production-pass / session-13 direct-child | `s31-v1-image.png` is broad-run historical evidence. `session-13` generated `s31-v2` directly in the child session without Agent/SubAgent; brief, prompt, and image exist; review mode `none`. User accepted the image. | | `s06` | accepted / mechanism pass | `s06-v2` prompt/image/return exist from clean `session-04` continuation with `Agent S-06` `019ef886-8797-7243-9427-d45065fcfc47`, no review file. Prompt records `gpt-image-2` Advisor, style module A, and visible back/walking character refs. User accepted the image. | | `s21` | generated / user accepted | `session-02` remains fallback evidence only. Codex thread-level child `019ef7d2-1a7f-7a90-8c04-0098239dfe0f` resolved `s21-v3`, started separate `Agent S-21` thread `019ef7d6-3ff6-7213-b68c-fd646e39f1d2`, wrote prompt/image/return, and correctly suppressed review. User accepted the image. | | `s14` | accepted / mechanism defect | `s14-v2` prompt/image/return exist, no review file. Prompt records `gpt-image-2` Advisor, visible character refs, and style B. User accepted the image. Mechanism is not clean: duplicate `Agent S-14` threads `019ef803-95d7-7411-8d3c-c7f2f68a32e7` and `019ef803-328a-7e40-a0b0-0333c6b5fc6c` targeted the same v2 files; return/image source belongs to the duplicate thread folder while child handoff records the other S-14 id. | @@ -198,7 +221,8 @@ - Duplicate-worker prevention remains a historical defect from `session-03`, but `session-04` did not reproduce it in a two-step sequential continuation. - R4 formal-run produced visually accepted S-02/S-03/S-04 outputs but exposed a brief-ownership mechanism issue. R5 Package 1 (`S-07`) and successful Package 2 outputs (`S-08/s08-v3`, `S-09/s09-v2`) passed the child-owned brief/packet gate in the same clean visible child session. - `s08-v2` remains incomplete transport-failure evidence and should not be backfilled or treated as an accepted output. -- `session-07` proved a direct 3-slide package can complete with accepted images and 157K observed context; the next test is a user-authorized 4-slide package to measure feasibility beyond the default 3-slide cap. +- `session-07` proved an Agent-based 3-slide package can complete with accepted images and 157K observed context. - R6 Phase 1 proved the callback path can wake this parent workflow after a production-shaped single-slide repair. `s16-v3` was accepted by the user. - R6 Phase 2 retry proved callback-only completion for a 4-shot package at the mechanical workflow level, and the user accepted `s20-v3`, `s22-v3`, `s23-v3`, and `s24-v2`. -- Child-session launch rule is closed for this repo default: use clean user-visible Codex child/session threads for brief/orchestration work, continue the same thread only inside the approved scope, and use internal SubAgent / one-shot slide workers for each slide. +- R7 proved production-light simplification: `session-12` completed an Agent-backed 3-slide package at 25m48s / 134K tokens; `session-13` completed a direct-child/no-Agent 3-slide package at 10m32s / 114K tokens. +- Child-session launch rule is closed for this repo default: use clean named user-visible Codex child/session threads (`子会话：...`) for production-light, let the child directly write brief/prompt/image/status/handoff for up to 3 slides, and reserve Agent/SubAgent for test/audit, repair, explicit review, exception, or isolated single-slide work. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/task-plan.md b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/task-plan.md index 286526f..783ccca 100644 --- a/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/task-plan.md +++ b/projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/task-plan.md @@ -5,7 +5,7 @@ - Task id: `2026-06-24-character-consistency` - Worker profile: `video-slide-production` - Running mode: `prompt and generate` -- Status: R6 Phase 2 retry callback completed; `session-11` completion gate pass; user accepted `S-20/S-22/S-23/S-24` +- Status: R7 production-light route accepted; `session-13` direct-child/no-Agent package for `S-29/S-30/S-31` is the formal production default - Child session reuse test: passed for `session-04` continuation (`S-33` then `S-06` in the same clean child thread) - Deprecated child session id: `019ef77d-552b-7343-abe4-88921b35d7ea` (do not reuse) - Session-02 child-session id: `019ef7ca-413a-7662-9633-7fb126b7ef35` @@ -47,8 +47,49 @@ - Session-11 parent callback target: `019ef9b5-cc39-7a42-91b1-f9b0852dc95e` - Session-11 scope: `S-20`, `S-22`, `S-23`, `S-24` - Session-11 result: completion gate `pass`; generated `s20-v3`, `s22-v3`, `s23-v3`, `s24-v2`; user accepted all four images +- Session-12 metric: user observed 25m48s / 134K tokens for 3 slides with Agent still used. +- Session-13 metric: user observed 10m32s / 114K tokens for 3 slides with direct child session and no Agent/SubAgent. - Artifact policy: append-only next slide version; overwrite disabled by default +## R7 Production-Light Active Scope + +- Line: `production-light` +- Worker profile: `video-slide-production` +- Running mode: `prompt and generate` +- Review mode: `none` +- Scope: `S-26/S-27/S-28` +- Expected versions from current slide folders: `s26-v2`, `s27-v2`, `s28-v2` +- Child session: `session-12` clean visible Codex child thread `019efbb4-b366-7de3-8e02-01183b3d0b13` +- Parent callback target: `019efb9f-ece2-78d2-ba6a-4ca1dcfe3244` +- Write policy: append-only; overwrite allowed: no +- Production rule: generated images are production-pass outputs by default; no per-image user review gate +- Parent consumption rule: read compact child handoff after callback; inspect prompt / return / image details only on `blocked`, `failed`, `warning`, artifact collision, duplicate worker, or explicit user audit request +- Status: complete; callback received; completion gate `pass`; outputs recorded as production-pass by task rule; user observed 25m48s / 134K tokens + +## Direct-Child Production-Light Test Scope + +- Line: `production-light` +- Topology: `main session -> child session directly`; no Agent / SubAgent +- Running mode: `prompt and generate` +- Review mode: `none` +- Scope: `S-29/S-30/S-31` +- Expected versions from current slide folders: `s29-v2`, `s30-v2`, `s31-v2` +- Child session: `session-13` clean visible Codex child thread `019efbd8-f75d-7512-9952-aa3528fde581` +- Parent callback target: `019efb9f-ece2-78d2-ba6a-4ca1dcfe3244` +- Write policy: append-only; overwrite allowed: no +- Test purpose: measure real time and context/token use when the child session directly writes brief, prompt, image, slides table, and compact handoff without Agent, packet, return receipt, review, or final verification pass. +- Status: complete; callback received; outputs recorded as production-pass by task rule; user accepted all three images and observed 10m32s / 114K tokens. This direct-child topology is now the formal production default. + +## Formal Production Default + +- Project phase source: `project.md` records `Workflow phase: production-light`. +- Default child session title: `子会话：`. +- Default topology: main session -> named child session directly. +- Default child scope: up to 3 slides unless user explicitly overrides. +- Default artifacts: `sNN-vN-brief.md`, `sNN-vN-prompt.md`, `sNN-vN-image.png`, `slides.md` status update, and compact child handoff. +- Default omissions: no review file, no Agent/SubAgent, no packet/return receipt, no routine final verification pass. +- Retained guard: scan target slide folder before writing, choose next unused version, and keep overwrite disabled. + ## User Instructions Latest third-round test instruction: @@ -102,10 +143,12 @@ s12 s13 ``` -Authorized next: +Completed R7 production-light trials: ```text -None; wait for the next user production decision after R6 Phase 2 retry acceptance. +S-26/S-27/S-28 as session-12, prompt and generate, review: none, Agent-backed production-light. +S-29/S-30/S-31 as session-13, prompt and generate, review: none, direct child/no-Agent. +No active authorized scope; wait for next production decision. ``` ## Child Session Split @@ -121,6 +164,8 @@ session-08: clean child thread `019ef969-66e7-7c52-a437-045371668d41`; S-15/S-16 session-09: clean child thread `019ef9bc-fe96-75b3-b902-5d5ef95feca3`; R6 Phase 1 callback-only S-16 repair generated `s16-v3`; completion gate pass, user accepted image session-10: clean child thread `019ef9d0-4914-7ea0-8935-bfd9bfc1aad4`; R6 Phase 2 first 4-shot package for S-19/S-20/S-22/S-23; paused by user after S-20 worker boundary violation; s19-v2 accepted, remaining v2 items incomplete evidence session-11: clean child thread `019ef9f8-8c34-77a3-82af-9a25beb2c723`; R6 Phase 2 retry package for S-20/S-22/S-23/S-24; callback received, completion gate pass, user accepted all four images +session-12: clean child thread `019efbb4-b366-7de3-8e02-01183b3d0b13`; R7 Agent-backed production-light package for S-26/S-27/S-28; user accepted all three images +session-13: clean child thread `019efbd8-f75d-7512-9952-aa3528fde581`; R7 direct-child/no-Agent production-light package for S-29/S-30/S-31; user accepted all three images; formal production default ``` Current `session-03` continuation generated `s14-v2` and `s25-v2` after user accepted `s21-v3-image.png`; user then accepted both images and closed the test. The mechanism result is not a clean pass because two S-14 workers were created for one slide/version. @@ -202,7 +247,7 @@ Session-04 result: - write the prompt/spec through the `gpt-image-2` Skill Advisor workflow; - load the real reference image into the conversation context before generation; - call host `image_gen` with the loaded image assigned a role such as `Image 1`; - - record the input path, output path, and reference-path caveat in the review or return. + - record the input path, output path, and reference-path caveat in the compact handoff or task status; only Agent-backed test/audit or repair runs need a return. ## Rules @@ -212,25 +257,37 @@ Session-04 result: - Do not treat old broad-run images as accepted finals. - Do not generate visible-character identity shots from text-only prompts. - Do not use archived character or style references unless explicitly reopened. -- Each slide agent handles exactly one slide. -- Child sessions and slide agents run sequentially. -- One slide/version must have exactly one owning slide-agent. A child session must not create two workers with the same packet and same output targets; prompt/image/return ownership must stay with the same worker id. -- Child-session capacity for the next formal-run simulation returns to the workflow default: maximum 3 slides per child session unless the task plan explicitly narrows it. -- Child session carrier rule: brief/orchestration child sessions should be Codex child/session threads visible to the user and recorded by thread id. Slide agents inside those child sessions should be internal SubAgent / one-shot workers, not long-lived reusable child sessions. +- When a slide agent is used, it handles exactly one slide. +- Child sessions and any slide agents run sequentially. +- In Agent-backed test/audit, repair, or exception runs, one slide/version must have exactly one owning slide-agent. A child session must not create two workers with the same packet and same output targets; prompt/image/return ownership must stay with the same worker id. +- Child-session capacity for formal production-light defaults to maximum 3 slides per child session unless the task plan explicitly changes it. +- Child session carrier rule: production-light uses clean named Codex child/session threads visible to the user and recorded by thread id. The child directly writes brief/prompt/image/status/handoff by default. Slide agents are reserved for test/audit, repair, explicit review, exception, or isolated single-slide work. - Production child sessions must start as clean user-visible threads. Do not use forked or inherited-context threads unless the user explicitly requests inherited context. -- Main-session production wait rule: choose the wait schedule by package size. A 1-slide package sleeps 10 minutes before the first status check and polls every 1 minute up to 15 minutes total; a 2-slide package sleeps 20 minutes before the first status check and polls every 1 minute up to 30 minutes total; a 3-slide package sleeps 30 minutes before the first status check and polls every 1 minute up to 45 minutes total. Split larger packages unless a task plan explicitly defines a custom wait schedule. -- Parent-facing heartbeat/status updates during routine waiting default to every 5 minutes or material state changes; these updates are separate from child-session status checks and should not trigger extra thread inspection. -- Child-session completion gate is mandatory. Instruction / visual review is optional and runs only when requested. -- Slide Agents are one-shot and become terminal / closed after return; do not continue or reuse them. +- Main-session production wait rule: callback-only by default. The parent dispatches, records the callback target, and stops routine polling. Time-based polling schedules are fallback only when callback tooling is unavailable or the user explicitly requests polling for a named run. +- Parent-facing heartbeat/status updates should not trigger extra child-thread inspection in routine production. +- Production-light closeout is minimal: version scan, no overwrite, image paths/status recorded, review absent when `review: none`, compact handoff before callback. The fuller completion gate is for test/audit, repair, explicit review, or exception scopes. +- Slide Agents, when used, are one-shot and become terminal / closed after return; do not continue or reuse them. - Main session records the new child session id so later slides can be sent to the same child session after user approval. - Resolve the slide artifact version before each slide run by scanning existing `slides/sNN/sNN-v*-*` files. - Do not overwrite existing slide artifacts. If any planned target file already exists, stop and report an artifact collision. - Do not backfill missing prompt or brief files for old broad-run images as if they were original generation inputs. -- For no-review workflow tests, a `sNN-vN-review.md` file should not be produced unless the task plan is changed. +- In production-light, `sNN-vN-review.md` should not be produced unless the task plan is changed. ## Expected Outputs -Task records: +Current production-light expected outputs: + +```text +projects/2026-06-23-在路上/slides/sNN/sNN-vN-brief.md +projects/2026-06-23-在路上/slides/sNN/sNN-vN-prompt.md +projects/2026-06-23-在路上/slides/sNN/sNN-vN-image.png +projects/2026-06-23-在路上/slides/slides.md update +projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-NN/child-session-handoff.md +``` + +Do not expect `sNN-vN-review.md`, slide-agent packet files, or slide-agent return files in production-light unless the task explicitly overrides the default. + +Legacy session-04 mechanism-test task records: ```text projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/task-plan.md @@ -246,7 +303,7 @@ projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-04/ projects/2026-06-23-在路上/tasks/2026-06-24-character-consistency/session-04/returns/s06-v2-slide-agent-return.md ``` -Slide records: +Legacy session-04 slide records: ```text projects/2026-06-23-在路上/slides/s33/s33-v2-brief.md @@ -257,7 +314,9 @@ projects/2026-06-23-在路上/slides/s06/s06-v2-prompt.md projects/2026-06-23-在路上/slides/s06/s06-v2-image.png ``` -## User Review / Sampling Plan +## Historical User Review / Sampling Plan + +These reporting gates are closed historical test notes. Production-light now defaults to no per-image user review; user review happens only by explicit sampling, repair request, or test/audit override. After `s21` generation, stop and report: @@ -302,4 +361,4 @@ This reporting gate is closed for `session-04`: the user confirmed the mechanism - Host image generation cannot be called. - Generated image cannot be copied into the project slide folder. - Any planned output path already exists and overwrite is not explicitly authorized. -- The result clearly fails the side/back dynamic character-continuity purpose. +- The child reports `blocked`, `failed`, or `warning`, or a generated result is later judged unusable and needs an append-only retry. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/production-run-status.md b/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/production-run-status.md new file mode 100644 index 0000000..8775519 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/production-run-status.md @@ -0,0 +1,60 @@ +# Production Run Status + +## Status + +- Project: `projects/2026-06-23-在路上` +- Lifecycle phase: production +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Package size: 8-slide validation override for this task only +- User confirmation: received; production-stage default is no per-image human review +- Current status: accepted; default package rule promoted + +## Scope + +- Total target units: S-38, S-39, S-40, S-41, S-42, S-43, S-44, S-45 +- Completed units: S-38, S-39, S-40, S-41, S-42, S-43, S-44, S-45 +- Active child session: none; child session `019efc48-d4af-7972-8646-6d04778dfb2c` completed +- Next cursor: remaining not-started scope `S-48/S-49/S-50` + +## Callback Target + +- Parent thread: 019efc02-1a26-7e73-9bb6-c1d540101d52 +- Callback mode: child writes handoff first, then sends callback as final action + +## Latest Callback + +- Status: complete +- Topology: direct-child +- Review mode: none +- Resolved versions: S-38 -> `s38-v1`; S-39 -> `s39-v1`; S-40 -> `s40-v1`; S-41 -> `s41-v1`; S-42 -> `s42-v1`; S-43 -> `s43-v1`; S-44 -> `s44-v1`; S-45 -> `s45-v1` +- Handoff path: `projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/session-01/child-session-handoff.md` +- Image paths: `projects/2026-06-23-在路上/slides/s38/s38-v1-image.png`; `projects/2026-06-23-在路上/slides/s39/s39-v1-image.png`; `projects/2026-06-23-在路上/slides/s40/s40-v1-image.png`; `projects/2026-06-23-在路上/slides/s41/s41-v1-image.png`; `projects/2026-06-23-在路上/slides/s42/s42-v1-image.png`; `projects/2026-06-23-在路上/slides/s43/s43-v1-image.png`; `projects/2026-06-23-在路上/slides/s44/s44-v1-image.png`; `projects/2026-06-23-在路上/slides/s45/s45-v1-image.png` +- Problems: none +- Needs user: none +- Parent verification: expected files present, exact review files absent, slide rows updated, and child thread tail shows no project file writes or checks after the callback/final child response. +- User acceptance/rule promotion: user promoted this 8-slide validation to production-pass and reported 174K tokens / 67% context / 38m37s elapsed. + +## Compression Reload Surface + +If the parent is compressed before or after callback, reload: + +1. `AGENTS.md` +2. `handoff/2026-06-25-on-the-road-formal-production-handoff.md` +3. `docs/workflows/project-lifecycle.md` +4. `docs/workflows/slide-task-orchestration.md` +5. `docs/workflows/slide-task-templates.md` +6. `projects/2026-06-23-在路上/project.md` +7. `projects/2026-06-23-在路上/execution-plan.md` +8. `projects/2026-06-23-在路上/slides/slides.md` +9. `projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/production-run-status.md` +10. `projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/task-log.md` + +Do not read prompt/image details unless the callback reports `blocked`, `failed`, `warning`, artifact collision, or the user asks for inspection. + +## Stop Conditions + +- Stop on `blocked`, `failed`, `warning`, artifact collision, missing reference, tool failure, or ambiguous scope/status. +- Stop after clean completion only if the handoff reports a problem or the task explicitly asks for human review. +- Default ordinary production-light package size is now up to 8 slides. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/session-01/child-session-handoff.md b/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/session-01/child-session-handoff.md new file mode 100644 index 0000000..565df78 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/session-01/child-session-handoff.md @@ -0,0 +1,21 @@ +# Child Session 01 Handoff + +- Status: complete +- Topology: direct-child +- Callback: ready-to-send-as-final-action +- Review mode: none +- Resolved versions: S-38 -> `s38-v1`; S-39 -> `s39-v1`; S-40 -> `s40-v1`; S-41 -> `s41-v1`; S-42 -> `s42-v1`; S-43 -> `s43-v1`; S-44 -> `s44-v1`; S-45 -> `s45-v1` +- Image paths: + - `projects/2026-06-23-在路上/slides/s38/s38-v1-image.png` + - `projects/2026-06-23-在路上/slides/s39/s39-v1-image.png` + - `projects/2026-06-23-在路上/slides/s40/s40-v1-image.png` + - `projects/2026-06-23-在路上/slides/s41/s41-v1-image.png` + - `projects/2026-06-23-在路上/slides/s42/s42-v1-image.png` + - `projects/2026-06-23-在路上/slides/s43/s43-v1-image.png` + - `projects/2026-06-23-在路上/slides/s44/s44-v1-image.png` + - `projects/2026-06-23-在路上/slides/s45/s45-v1-image.png` +- Problems: none reported by the child session +- Needs user: visual confirmation required +- Observed elapsed/context: not visible in this child-session interface +- Narrow check: expected `brief.md`, `prompt.md`, and `image.png` files exist for S-38 through S-45; exact `sNN-v1-review.md` targets are absent as required +- Parent summary: S-38 through S-45 were generated in the 8-slide validation override using the direct-child production-light path, no Agent/SubAgent, review `none`, append-only `v1` versions, `gpt-image-2` Advisor / host-native prompt workflow, visible references loaded before generation when needed, and `slides/slides.md` updated with `generated / validation pending`. Do not mark this validation passed or change package defaults until user visual confirmation. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/task-log.md b/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/task-log.md new file mode 100644 index 0000000..b624e03 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/task-log.md @@ -0,0 +1,34 @@ +# Task Log + +## Status + +- Current status: accepted; default package rule promoted +- Running mode: prompt and generate +- Active line: production-light +- Review mode: none +- Validation package size: 8 slides; promoted to ordinary production-light default after acceptance +- User confirmation: received in parent thread; no per-image human review required in production-light +- Write policy: append-only; overwrite allowed: no +- Callback final-action rule: required + +## Dispatch / Callback Receipts + +- 2026-06-25 dispatch: child thread `019efc48-d4af-7972-8646-6d04778dfb2c`, scope `S-38/S-39/S-40/S-41/S-42/S-43/S-44/S-45`, line `production-light`, topology `direct-child/no-Agent`, review `none`, callback target `019efc02-1a26-7e73-9bb6-c1d540101d52`. +- 2026-06-25 callback: status `complete`; topology `direct-child`; review `none`; resolved versions `S-38:s38-v1`, `S-39:s39-v1`, `S-40:s40-v1`, `S-41:s41-v1`, `S-42:s42-v1`, `S-43:s43-v1`, `S-44:s44-v1`, `S-45:s45-v1`; handoff `projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/session-01/child-session-handoff.md`; problems `none`; initial callback requested visual confirmation before the production-stage default was clarified. +- 2026-06-25 parent verification: child handoff exists; expected brief/prompt/image files exist; exact review files are absent; `slides/slides.md` rows were generated, then promoted to `production-pass / 8-slide validation accepted`; child thread tail shows no project file writes or checks after the callback/final child response. +- 2026-06-25 user acceptance/rule promotion: user clarified the project is in production, images are default pass without human review, and the 8-slide package should become the global/local default. User reported package metrics: 174K tokens, 67% context, 38m37s elapsed. + +## Slide Status Summary + +- `S-38`: production-pass / 8-slide validation accepted; files `s38-v1-brief.md`, `s38-v1-prompt.md`, `s38-v1-image.png`. +- `S-39`: production-pass / 8-slide validation accepted; files `s39-v1-brief.md`, `s39-v1-prompt.md`, `s39-v1-image.png`. +- `S-40`: production-pass / 8-slide validation accepted; files `s40-v1-brief.md`, `s40-v1-prompt.md`, `s40-v1-image.png`. +- `S-41`: production-pass / 8-slide validation accepted; files `s41-v1-brief.md`, `s41-v1-prompt.md`, `s41-v1-image.png`. +- `S-42`: production-pass / 8-slide validation accepted; files `s42-v1-brief.md`, `s42-v1-prompt.md`, `s42-v1-image.png`. +- `S-43`: production-pass / 8-slide validation accepted; files `s43-v1-brief.md`, `s43-v1-prompt.md`, `s43-v1-image.png`. +- `S-44`: production-pass / 8-slide validation accepted; files `s44-v1-brief.md`, `s44-v1-prompt.md`, `s44-v1-image.png`. +- `S-45`: production-pass / 8-slide validation accepted; files `s45-v1-brief.md`, `s45-v1-prompt.md`, `s45-v1-image.png`. + +## Open Decisions + +- No open decision. Default ordinary production-light package size is now up to 8 slides. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/task-plan.md b/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/task-plan.md new file mode 100644 index 0000000..0dbb461 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-8slide-compression-validation/task-plan.md @@ -0,0 +1,78 @@ +# 8-Slide Compression Validation + +## Active Scope + +- Line: production-light +- Worker profile: video-slide-production +- Running mode: prompt and generate +- Review mode: none +- Slides: S-38, S-39, S-40, S-41, S-42, S-43, S-44, S-45 +- Child session: `019efc48-d4af-7972-8646-6d04778dfb2c` +- Callback target: 019efc02-1a26-7e73-9bb6-c1d540101d52 +- Write policy: append-only; overwrite allowed: no +- Status: accepted; default package rule promoted + +## Context To Use + +- `AGENTS.md` +- `VIDEO_WORKBENCH.md` +- `docs/workflows/project-lifecycle.md` +- `docs/workflows/slide-task-orchestration.md` +- `docs/workflows/slide-task-templates.md` +- `handoff/2026-06-25-on-the-road-formal-production-handoff.md` +- `projects/2026-06-23-在路上/project.md` +- `projects/2026-06-23-在路上/execution-plan.md` +- `projects/2026-06-23-在路上/slides/slides.md` +- `projects/2026-06-23-在路上/visual-system/visual-system.md` +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md` +- Assigned slide folders under `projects/2026-06-23-在路上/slides/sNN/` + +## Validation Override + +- Package size: 8 slides for this task only. +- This exceeds the current ordinary production default of 5 slides and does not change that default. +- Validation goals: + - Confirm callback is the child session's final action after all allowed writes/checks/handoff work. + - Observe whether child context still has margin after an 8-slide package. + - Provide a compression-safe parent continuation surface after callback. +- User confirmation received in parent thread: production-stage default is no per-image human review; this 8-slide validation is accepted and promoted to the ordinary production-light package rule. +- Observed production metrics from user: 174K tokens, 67% context, 38m37s elapsed. +- Default ordinary production-light package size is now up to 8 slides. +- Parent verification after callback: + - Child handoff exists and reports `complete`. + - Expected brief, prompt, and image files exist for S-38 through S-45. + - Exact `sNN-v1-review.md` targets are absent. + - `slides/slides.md` recorded all eight rows as generated, then the parent promoted them to `production-pass / 8-slide validation accepted` after user rule clarification. + - Child thread tail shows no project file writes or checks after the callback/final child response. + +## Expected Versions + +- S-38: expected `s38-v1` if files are unchanged. +- S-39: expected `s39-v1` if files are unchanged. +- S-40: expected `s40-v1` if files are unchanged. +- S-41: expected `s41-v1` if files are unchanged. +- S-42: expected `s42-v1` if files are unchanged. +- S-43: expected `s43-v1` if files are unchanged. +- S-44: expected `s44-v1` if files are unchanged. +- S-45: expected `s45-v1` if files are unchanged. + +The child session must scan each target slide folder before writing. These expected versions are not authority. + +## Callback Final-Action Rule + +- The child must complete all allowed project writes, `slides/slides.md` updates, narrow required checks, and `child-session-handoff.md` before callback. +- The callback is the final child action. +- After a successful callback, the child must not write, verify, inspect, or update project files. +- The handoff should say `Callback: ready-to-send-as-final-action`; the parent callback itself is the sent receipt. + +## Stop Conditions + +- Any target path already exists for the resolved version. +- Required visual-system or reference files cannot be read. +- The active generation path cannot load a required visible-context image reference for an identity-relevant slide. +- The child cannot use `gpt-image-2` Advisor / host-native workflow. +- Tooling tries to call MiniMax MCP, CCPE, GPT V2, web search, or external vision/review services. +- The child reports `blocked`, `failed`, or `warning`. +- The child finishes cleanly; parent must pause for user visual confirmation before changing package-size defaults. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/production-run-status.md b/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/production-run-status.md new file mode 100644 index 0000000..285d487 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/production-run-status.md @@ -0,0 +1,57 @@ +# Production Run Status + +## Status + +- Project: `projects/2026-06-23-在路上` +- Lifecycle phase: assembly-closeout +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Package size: 3 slides; within 8-slide production default +- User confirmation: production-stage default; no per-image human review required +- Current status: complete; parent verified + +## Scope + +- Total target units: S-48, S-49, S-50 +- Completed units: S-48, S-49, S-50 +- Active child session: none; child session `019efc78-e0f1-7601-aaec-bab4f3e6f99d` completed +- Next cursor: none; shot-image production complete for current 51-shot table + +## Callback Target + +- Parent thread: 019efc02-1a26-7e73-9bb6-c1d540101d52 +- Callback mode: child writes handoff first, then sends callback as final action + +## Latest Callback + +- Status: complete +- Topology: direct-child +- Resolved versions: `s48-v1`, `s49-v1`, `s50-v1` +- Handoff path: `projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/session-01/child-session-handoff.md` +- Image paths: `projects/2026-06-23-在路上/slides/s48/s48-v1-image.png`; `projects/2026-06-23-在路上/slides/s49/s49-v1-image.png`; `projects/2026-06-23-在路上/slides/s50/s50-v1-image.png` +- Problems: none +- Needs user: none +- Parent verification: expected files present, exact review files absent, slide rows updated, and child thread tail shows no project file writes or checks after callback. + +## Compression Reload Surface + +If the parent is compressed before or after callback, reload: + +1. `AGENTS.md` +2. `handoff/2026-06-25-on-the-road-formal-production-handoff.md` +3. `docs/workflows/project-lifecycle.md` +4. `docs/workflows/slide-task-orchestration.md` +5. `docs/workflows/slide-task-templates.md` +6. `projects/2026-06-23-在路上/project.md` +7. `projects/2026-06-23-在路上/execution-plan.md` +8. `projects/2026-06-23-在路上/slides/slides.md` +9. `projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/production-run-status.md` +10. `projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/task-log.md` + +Do not read prompt/image details unless the callback reports `blocked`, `failed`, `warning`, artifact collision, or the user asks for inspection. + +## Stop Conditions + +- Stop on `blocked`, `failed`, `warning`, artifact collision, missing reference, tool failure, or ambiguous scope/status. +- Continue normally when callback reports `complete` with no problems. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/session-01/child-session-handoff.md b/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/session-01/child-session-handoff.md new file mode 100644 index 0000000..2171ce0 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/session-01/child-session-handoff.md @@ -0,0 +1,18 @@ +# Child Session 01 Handoff + +- Status: complete +- Topology: direct-child +- Callback: ready-to-send-as-final-action +- Review mode: none +- Resolved versions: `S-48 -> s48-v1`; `S-49 -> s49-v1`; `S-50 -> s50-v1` +- Image paths: + - `projects/2026-06-23-在路上/slides/s48/s48-v1-image.png` + - `projects/2026-06-23-在路上/slides/s49/s49-v1-image.png` + - `projects/2026-06-23-在路上/slides/s50/s50-v1-image.png` +- Problems: none +- Needs user: none +- Observed elapsed/context: not visible to child session + +## Parent Summary + +Final production package `S-48/S-49/S-50` completed through direct child production-light. The child wrote per-slide briefs and prompts, used `gpt-image-2` Advisor / host-native Mode B workflow, loaded visible references before generation, copied host-generated images into the project slide folders, updated `slides/slides.md`, `task-log.md`, and `production-run-status.md`, and kept review mode `none` with no `sNN-vN-review.md` files. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/task-log.md b/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/task-log.md new file mode 100644 index 0000000..7c0f226 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/task-log.md @@ -0,0 +1,34 @@ +# Task Log + +## Status + +- Current status: complete; parent verified +- Running mode: prompt and generate +- Active line: production-light +- Review mode: none +- Package size: 3 slides; within 8-slide production default +- User confirmation: production-stage default; no per-image human review required +- Write policy: append-only; overwrite allowed: no +- Callback final-action rule: required + +## Dispatch / Callback Receipts + +- 2026-06-25 dispatch: child thread `019efc78-e0f1-7601-aaec-bab4f3e6f99d`, scope `S-48/S-49/S-50`, line `production-light`, topology `direct-child/no-Agent`, review `none`, callback target `019efc02-1a26-7e73-9bb6-c1d540101d52`. +- 2026-06-25 10:02 child ready: complete, topology `direct-child`, versions `s48-v1/s49-v1/s50-v1`, handoff `projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/session-01/child-session-handoff.md`, needs_user `none`. +- 2026-06-25 parent verification: handoff exists and reports `complete`; expected brief/prompt/image files exist for S-48/S-49/S-50; exact review files are absent; `slides/slides.md` records all three rows as `production-pass / final production direct-child`; child thread tail shows no project file writes or checks after callback. + +## Slide Status Summary + +- `S-48`: production-pass / final production direct-child; files `s48-v1-brief.md`, `s48-v1-prompt.md`, `s48-v1-image.png`; review file absent by instruction. +- `S-49`: production-pass / final production direct-child; files `s49-v1-brief.md`, `s49-v1-prompt.md`, `s49-v1-image.png`; review file absent by instruction. +- `S-50`: production-pass / final production direct-child; files `s50-v1-brief.md`, `s50-v1-prompt.md`, `s50-v1-image.png`; review file absent by instruction. + +## Image Paths + +- `projects/2026-06-23-在路上/slides/s48/s48-v1-image.png` +- `projects/2026-06-23-在路上/slides/s49/s49-v1-image.png` +- `projects/2026-06-23-在路上/slides/s50/s50-v1-image.png` + +## Open Decisions + +- None. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/task-plan.md b/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/task-plan.md new file mode 100644 index 0000000..c9d22a8 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-final-production-s48-s50/task-plan.md @@ -0,0 +1,60 @@ +# Final Production Package S48-S50 + +## Active Scope + +- Line: production-light +- Worker profile: video-slide-production +- Running mode: prompt and generate +- Review mode: none +- Slides: S-48, S-49, S-50 +- Child session: `019efc78-e0f1-7601-aaec-bab4f3e6f99d` +- Callback target: 019efc02-1a26-7e73-9bb6-c1d540101d52 +- Write policy: append-only; overwrite allowed: no +- Status: complete; parent verified + +## Context To Use + +- `AGENTS.md` +- `VIDEO_WORKBENCH.md` +- `docs/workflows/project-lifecycle.md` +- `docs/workflows/slide-task-orchestration.md` +- `docs/workflows/slide-task-templates.md` +- `handoff/2026-06-25-on-the-road-formal-production-handoff.md` +- `projects/2026-06-23-在路上/project.md` +- `projects/2026-06-23-在路上/execution-plan.md` +- `projects/2026-06-23-在路上/slides/slides.md` +- `projects/2026-06-23-在路上/visual-system/visual-system.md` +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md` +- Assigned slide folders under `projects/2026-06-23-在路上/slides/s48`, `s49`, and `s50` + +Do not read old session directories or old handoffs unless a specific inconsistency appears. + +## Expected Versions + +- S-48: expected `s48-v1` if files are unchanged. +- S-49: expected `s49-v1` if files are unchanged. +- S-50: expected `s50-v1` if files are unchanged. + +The child session must scan each target slide folder before writing. These expected versions are not authority. + +## Production Rules + +- Production phase is active; generated images are default production-pass outputs. +- No per-image human review is required for this package. +- Do not create review files. +- Do not launch Agent/SubAgent. +- Do not create packet or return files. +- Do not run a routine final completeness audit. +- Complete all allowed writes, narrow checks, and `child-session-handoff.md` before callback. +- Callback is the final child action; after successful callback, do not write, verify, inspect, or update project files. + +## Stop Conditions + +- Any target path already exists for the resolved version. +- Required visual-system or reference files cannot be read. +- The active generation path cannot load a required visible-context image reference for an identity-relevant slide. +- The child cannot use `gpt-image-2` Advisor / host-native workflow. +- Tooling tries to call MiniMax MCP, CCPE, GPT V2, web search, or external vision/review services. +- The child reports `blocked`, `failed`, `warning`, artifact collision, or ambiguous scope/status. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/production-run-status.md b/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/production-run-status.md new file mode 100644 index 0000000..d995d6a --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/production-run-status.md @@ -0,0 +1,47 @@ +# Production Run Status + +## Status + +- Project: `projects/2026-06-23-在路上` +- Lifecycle phase: production +- Line: production-light +- Running mode: prompt and generate +- Review mode: none +- Package size: 5-slide validation override for this task only +- User confirmation: received; all five images accepted +- Current status: validation accepted by user; package-size default updated + +## Scope + +- Total target units: S-32, S-34, S-35, S-36, S-37 +- Completed units: S-32/s32-v2, S-34/s34-v2, S-35/s35-v1, S-36/s36-v1, S-37/s37-v1 (accepted by user) +- Active child session: `019efc24-96a4-7582-bb05-914b82a726d5` +- Next cursor: none in this validation scope; waiting for user confirmation + +## Callback Target + +- Parent thread: 019efc02-1a26-7e73-9bb6-c1d540101d52 +- Callback mode: child writes handoff first, then sends callback to parent + +## Latest Callback + +- Status: complete +- Topology: direct-child +- Resolved versions: S-32 -> s32-v2; S-34 -> s34-v2; S-35 -> s35-v1; S-36 -> s36-v1; S-37 -> s37-v1 +- Handoff path: `projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/session-01/child-session-handoff.md` +- Image paths: `projects/2026-06-23-在路上/slides/s32/s32-v2-image.png`; `projects/2026-06-23-在路上/slides/s34/s34-v2-image.png`; `projects/2026-06-23-在路上/slides/s35/s35-v1-image.png`; `projects/2026-06-23-在路上/slides/s36/s36-v1-image.png`; `projects/2026-06-23-在路上/slides/s37/s37-v1-image.png` +- Problems: none reported by child session +- Needs user: none for this validation; next session may start full unattended production + +## Capacity Result + +- User accepted all five images. +- User-reported child metrics: callback at about 152K tokens / 58% context; final observed usage after post-callback document/check work about 164K tokens / 63% context. +- Default package-size update: ordinary production-light child sessions may handle up to 5 slides. +- Reduction rule: use 1-3 slides for identity-heavy, repair-heavy, core-risk, or context-expensive scopes. +- Callback-order correction: future child sessions must treat callback as final action after all allowed writes/checks/handoff work. + +## Stop Conditions + +- Stop on `blocked`, `failed`, `warning`, artifact collision, missing reference, tool failure, or ambiguous scope/status. +- Stop only on blocker, failure, warning, artifact collision, missing reference, tool failure, ambiguous scope/status, or user-requested inspection. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/session-01/child-session-handoff.md b/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/session-01/child-session-handoff.md new file mode 100644 index 0000000..5d79128 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/session-01/child-session-handoff.md @@ -0,0 +1,23 @@ +# Child Session 01 Handoff + +- Status: complete +- Topology: direct-child +- Callback: sent to parent thread `019efc02-1a26-7e73-9bb6-c1d540101d52` +- Review mode: none +- Resolved versions: `S-32 -> s32-v2`; `S-34 -> s34-v2`; `S-35 -> s35-v1`; `S-36 -> s36-v1`; `S-37 -> s37-v1` +- Image paths: + - `projects/2026-06-23-在路上/slides/s32/s32-v2-image.png` + - `projects/2026-06-23-在路上/slides/s34/s34-v2-image.png` + - `projects/2026-06-23-在路上/slides/s35/s35-v1-image.png` + - `projects/2026-06-23-在路上/slides/s36/s36-v1-image.png` + - `projects/2026-06-23-在路上/slides/s37/s37-v1-image.png` +- Host source images: + - `C:\Users\wangq\.codex\generated_images\019efc24-96a4-7582-bb05-914b82a726d5\ig_05ee076cda100ad5016a3c758fc18c8191ad950eeaf7ce1fa1.png` -> `s32-v2-image.png` + - `C:\Users\wangq\.codex\generated_images\019efc24-96a4-7582-bb05-914b82a726d5\ig_056357067518abd7016a3c76014b408191b0df69d68dbbf3cd.png` -> `s34-v2-image.png` + - `C:\Users\wangq\.codex\generated_images\019efc24-96a4-7582-bb05-914b82a726d5\ig_0cb80a9e93cc7cdd016a3c766c2db0819198c9628145fda8fe.png` -> `s35-v1-image.png` + - `C:\Users\wangq\.codex\generated_images\019efc24-96a4-7582-bb05-914b82a726d5\ig_05dc979a2ac4dbb3016a3c77286a588191a487cd9d7cd35f1f.png` -> `s36-v1-image.png` + - `C:\Users\wangq\.codex\generated_images\019efc24-96a4-7582-bb05-914b82a726d5\ig_0809761295136eba016a3c78007db08191944085b291ed6fbb.png` -> `s37-v1-image.png` +- Problems: none reported by child session. +- Needs user: visual confirmation required before this capacity validation is treated as passed; do not change package-size defaults yet. +- Observed elapsed/context: not visible to this child session. +- Parent summary: Production-light validation package completed as direct-child work with no Agent/SubAgent, no packet/return files, and no review files. Briefs, prompts, images, and `slides/slides.md` were updated for `S-32`, `S-34`, `S-35`, `S-36`, and `S-37`. All updated slide rows use `generated / validation pending`, not `accepted` or `production-pass`. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/task-log.md b/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/task-log.md new file mode 100644 index 0000000..fb48b90 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/task-log.md @@ -0,0 +1,31 @@ +# Task Log + +## Status + +- Current status: validation accepted by user; package-size default updated +- Running mode: prompt and generate +- Active line: production-light +- Review mode: none +- Validation package size: 5 slides +- User confirmation: received; all five images accepted +- Write policy: append-only; overwrite allowed: no + +## Dispatch / Callback Receipts + +- 2026-06-25 dispatch: child thread `019efc24-96a4-7582-bb05-914b82a726d5`, scope `S-32/S-34/S-35/S-36/S-37`, line `production-light`, topology `direct-child/no-Agent`, review `none`, callback target `019efc02-1a26-7e73-9bb6-c1d540101d52`. +- 2026-06-25 callback: child thread `019efc24-96a4-7582-bb05-914b82a726d5`, status `complete`, topology `direct-child`, resolved versions `S-32 -> s32-v2`, `S-34 -> s34-v2`, `S-35 -> s35-v1`, `S-36 -> s36-v1`, `S-37 -> s37-v1`, handoff `session-01/child-session-handoff.md`, problems `none`, needs user `visual confirmation required`. +- 2026-06-25 user confirmation: all five images accepted. User-reported child metrics: callback at about 152K tokens / 58% context; after callback-side document/check work, final observed usage about 164K tokens / 63% context. +- 2026-06-25 mechanism note: child callback was not the final action; it updated the handoff and ran narrow checks after callback. Workflow docs were updated so future child sessions must complete all writes/checks before callback and treat callback as the final action. + +## Slide Status Summary + +- `S-32`: `s32-v2` accepted by user. +- `S-34`: `s34-v2` accepted by user. +- `S-35`: `s35-v1` accepted by user. +- `S-36`: `s36-v1` accepted by user. +- `S-37`: `s37-v1` accepted by user. + +## Open Decisions + +- Validation passed for quality and capacity. +- Future production child sessions should use up to 5 ordinary production slides by default, with callback as final action. diff --git a/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/task-plan.md b/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/task-plan.md new file mode 100644 index 0000000..f67bf76 --- /dev/null +++ b/projects/2026-06-23-在路上/tasks/2026-06-25-production-capacity-validation/task-plan.md @@ -0,0 +1,57 @@ +# Production Capacity Validation + +## Active Scope + +- Line: production-light +- Worker profile: video-slide-production +- Running mode: prompt and generate +- Review mode: none +- Slides: S-32, S-34, S-35, S-36, S-37 +- Child session: `019efc24-96a4-7582-bb05-914b82a726d5` +- Callback target: 019efc02-1a26-7e73-9bb6-c1d540101d52 +- Write policy: append-only; overwrite allowed: no +- Status: validation accepted by user; callback-order rule fixed in workflow docs + +## Context To Use + +- `AGENTS.md` +- `VIDEO_WORKBENCH.md` +- `docs/workflows/project-lifecycle.md` +- `docs/workflows/slide-task-orchestration.md` +- `docs/workflows/slide-task-templates.md` +- `handoff/2026-06-25-on-the-road-formal-production-handoff.md` +- `projects/2026-06-23-在路上/project.md` +- `projects/2026-06-23-在路上/execution-plan.md` +- `projects/2026-06-23-在路上/slides/slides.md` +- `projects/2026-06-23-在路上/visual-system/visual-system.md` +- `projects/2026-06-23-在路上/visual-system/style/style-prompt-master.md` +- `projects/2026-06-23-在路上/visual-system/characters/character-reference-registry.md` +- `projects/2026-06-23-在路上/intake/阶段五-镜头清单骨架.md` +- Assigned slide folders under `projects/2026-06-23-在路上/slides/sNN/` + +## Validation Override + +- Package size: 5 slides for this task only. +- Result: accepted by user on 2026-06-25. +- Global default update: ordinary production-light child sessions may now handle up to 5 slides; reduce to 1-3 for identity-heavy, repair-heavy, core-risk, or context-expensive scopes. +- Mechanism note: this validation exposed a callback-order defect. The child sent callback at about 152K tokens / 58% context, then updated handoff and ran narrow checks, ending around 164K tokens / 63% context. Workflow docs now require callback to be the child session's final action. + +## Expected Versions + +- S-32: expected `s32-v2` if files are unchanged. +- S-34: expected `s34-v2` if files are unchanged. +- S-35: expected `s35-v1` if files are unchanged. +- S-36: expected `s36-v1` if files are unchanged. +- S-37: expected `s37-v1` if files are unchanged. + +The child session must scan each target slide folder before writing. These expected versions are not authority. + +## Stop Conditions + +- Any target path already exists for the resolved version. +- Required visual-system or reference files cannot be read. +- The active generation path cannot load a required visible-context image reference for an identity-relevant slide. +- The child cannot use `gpt-image-2` Advisor / host-native workflow. +- Tooling tries to call MiniMax MCP, CCPE, GPT V2, web search, or external vision/review services. +- The child reports `blocked`, `failed`, or `warning`. +- The child finishes cleanly; parent must pause for user visual confirmation before changing package-size defaults.