diff --git a/AGENTS.md b/AGENTS.md
index fea5d86..2c52873 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -30,3 +30,27 @@ C:\Users\wangq\Documents\Codex\ccpe-system\runtimes\hybrid\ccpe-agent-consumer-s
 Do not hardcode old Agent, Lite, Skill, Runtime, or protocol paths as default CCPE entrypoints. Read narrower CCPE artifacts only when the consumer setup contract, a project-local registry, or the active task requires them.
 
 Formal CCPE participant output requires a project-local registry, an invocation packet, a real participant carrier, and returned output saved only after the participant returns. Main-session roleplay is not formal CCPE output.
+
+## Image Prompt Workflow
+
+All image-generation prompts in this repository must be produced through the `gpt-image-2` Skill workflow. Do not draft final image prompts directly from memory or from a derived project skeleton alone.
+
+Before writing any project-bound image prompt, run the Skill mode check and follow the resulting mode:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+- Mode A: use the Skill's Garden generation/edit scripts and save prompt/image traces under `garden-gpt-image-2/`.
+- Mode B: use the Skill as a prompt-engineering workflow, save the rendered prompt, then pass it to the host image tool only after user approval.
+- Mode C / Advisor: use the Skill as a high-quality prompt advisor, save the rendered prompt, and do not claim an image was generated.
+
+For every project-bound image prompt document:
+
+- Read the relevant `gpt-image-2` Skill template or reference file before drafting.
+- Record the Skill mode, template/reference path, and source files actually read.
+- Treat accepted `intake/` files and user-provided references as primary sources; use derived files such as `visual-system.md`, `execution-plan.md`, or `slides/slides.md` only as execution control summaries.
+- If a prompt was drafted without the Skill workflow, label it as a non-generation-ready draft and rebuild it through the Skill before any image generation.
+- When a local image outside the workspace is used, label its role explicitly as identity reference, edit target, or supporting reference, and do not proceed if the file cannot be read or approved.
+
+Project-bound image generation must remain approval-gated: save the prompt/spec first, wait for user approval, then execute generation only when explicitly requested.
diff --git a/VIDEO_WORKBENCH.md b/VIDEO_WORKBENCH.md
index 7ee7141..d1f3f99 100644
--- a/VIDEO_WORKBENCH.md
+++ b/VIDEO_WORKBENCH.md
@@ -176,7 +176,17 @@ Video generation from images remains manual or semi-manual for now because outpu
 
 Editing and publishing are human-controlled unless explicitly automated later.
 
-Codex is the default prompt owner for final image prompts in this workspace. It may use `gpt-image-2` Advisor or other image tools, but GPT V2 planning Markdown remains a source specification, not a final prompt.
+Codex is the default prompt owner for final image prompts in this workspace, but final prompts must be produced through the `gpt-image-2` Skill workflow rather than direct freehand drafting. GPT V2 planning Markdown remains a source specification, not a final prompt.
+
+For project-bound image prompts, Codex must:
+
+- run the `gpt-image-2` mode check before prompt work;
+- read the relevant Skill template or reference file;
+- save the rendered prompt/spec as a reviewable project file;
+- list the source files actually used;
+- wait for user approval before image generation.
+
+Direct prompts may exist only as clearly labeled scratch or non-generation-ready drafts. They must be rebuilt through the `gpt-image-2` Skill workflow before use.
 
 ## 12. Capability-Request Mode
 
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 0662913..f7ec28b 100644
--- a/projects/2026-06-23-在路上/execution-plan.md
+++ b/projects/2026-06-23-在路上/execution-plan.md
@@ -22,12 +22,12 @@ Immediate focus:
 
 | Target | Required Output | Status |
 | --- | --- | --- |
-| Character anchor board | Main 16:9 board on shallow gray-white background | pending |
-| Character transparent slices | Front full body, side full body, back full body, walking side, walking back | pending |
+| Character anchor board | Main 16:9 board on shallow gray-white background | accepted v2 |
+| Character transparent slices | Front full body, side full body, back full body, walking side, walking back | accepted v2-from-anchor set |
 | Character labels | Short identity, clothing, prop, and boundary labels | pending |
-| Style anchor board | 16:9 dark-gray board with A/B/C/D samples | pending |
-| C style retest | Dark muddy puddle / broken star reflection sample | pending |
-| Shot test package | S-01, S-10, S-46, S-47, S-51 | pending |
+| Style anchor board | 16:9 dark-gray board with A/B/C/D samples | accepted |
+| C style retest | Dark muddy puddle / broken star reflection sample | accepted |
+| Shot test package | S-01, S-10, S-46, S-47, S-51 | accepted |
 
 ## Small-Batch Strategy
 
@@ -53,15 +53,24 @@ Round 1 success means these are true:
 
 ## Current Round
 
-- Round: R0 skeleton.
+- Round: R2 five-shot trial preparation.
 - R0 completed: intake cleaned; project execution files initialized.
-- R1 next: character anchor board prompt/spec and first generation attempt.
+- R1 completed so far: v2 main character anchor accepted; identity lock approved for slice generation; five transparent slice candidates generated and rejected due identity drift from v2.
+- R1 completed: v2 main character anchor accepted; v2-from-anchor transparent slice set accepted.
+- R1 completed: C style retest accepted as `style-anchor-c-repair-v1.png`.
+- R1 completed: common style anchor board accepted as `style-anchor-board-v1.png`.
+- R2 completed: S-01 generated once and accepted by user.
+- R2 completed: user waived intermediate prompt review for S-10; S-10 generated once and accepted by user.
+- R2 completed: S-46 generated once and accepted by user.
+- R2 completed: S-47 generated once and accepted by user.
+- R2 completed: S-51 generated once and accepted by user.
+- R2 completed: five-shot visual trial generated and accepted by user.
 
 ## Active Batch
 
-No shot prompt batch is active yet.
+No multi-shot prompt batch is active yet.
 
-The first active batch will be created after character/style anchors have at least first accepted versions.
+No single-shot gate is active. The five-shot visual trial is accepted.
 
 ## Open Decisions
 
@@ -70,9 +79,9 @@ The first active batch will be created after character/style anchors have at lea
 | Character source face | Current stage does not use the real face; use a 50-year-old traveler based on the defined character system. |
 | 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 must be specially marked as priority repair. |
+| Style board scope | MVP one-page A/B/C/D board; C should be marked as repaired / accepted. |
 | Full shot production | Deferred until anchor tests are accepted. |
 
 ## Next Decision
 
-Proceed to the character anchor board first.
+Decide the lightweight broad-production flow for the remaining non-generated shots, including whether to skip per-shot review and generate missing shots in sequence.
diff --git a/projects/2026-06-23-在路上/project.md b/projects/2026-06-23-在路上/project.md
index 88d8665..0fbea34 100644
--- a/projects/2026-06-23-在路上/project.md
+++ b/projects/2026-06-23-在路上/project.md
@@ -38,7 +38,7 @@ The emotional center is not victory. It is continuing.
 | --- | --- | --- |
 | A: 风景主导型 | accepted | Main composition: world large, road long, person small. |
 | B: 高位观察型 | accepted | City edge and system pressure. |
-| C: 泥泞里的星空 | must be rebuilt | Soul image; needs darker night/pre-dawn puddle reflection, less sunrise, more broken starlight. |
+| 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
@@ -60,4 +60,11 @@ The emotional center is not victory. It is continuing.
 
 ## Next Action
 
-Start with the character anchor board. It is the identity basis for later shot prompts and should be accepted before broad shot generation.
+Decide the lightweight broad-production flow for the remaining non-generated shots.
+
+```text
+slides/slides.md
+execution-plan.md
+```
+
+The v2-from-anchor transparent character slice set is accepted. The C repair sample is accepted. The common style anchor board is accepted. The five-shot visual trial is accepted: S-01, S-10, S-46, S-47, and S-51 are generated and accepted by user.
diff --git a/projects/2026-06-23-在路上/slides/five-shot-trial-plan-v1.md b/projects/2026-06-23-在路上/slides/five-shot-trial-plan-v1.md
new file mode 100644
index 0000000..c0f36a2
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/five-shot-trial-plan-v1.md
@@ -0,0 +1,382 @@
+# Five-Shot Trial Plan V1
+
+## Status
+
+- Document status: user approved for S-01 brief creation.
+- Execution status: S-01, S-10, S-46, S-47, and S-51 generated and accepted by user.
+- Project: `在路上`.
+- Scope: first five-shot visual test package before any broad 51-shot production.
+- Trial shots: `S-01`, `S-10`, `S-46`, `S-47`, `S-51`.
+
+This file is not a final image-generation prompt.
+
+All later project-bound image prompts must be created through the `gpt-image-2` Skill workflow and must record:
+
+- Skill mode check;
+- Skill template/reference files read;
+- project sources actually used;
+- approved anchor assets used as references;
+- prompt approval state before generation.
+
+## Purpose
+
+The five-shot trial should answer one question:
+
+> Can the accepted character anchor and common style anchor actually survive real scene generation?
+
+The goal is not to produce five beautiful isolated images.
+
+The goal is to prove that:
+
+1. the traveler remains recognizable without becoming a portrait subject;
+2. A/B/C/D behave like one MV world;
+3. C stays repaired as muddy starlight, not warm sunrise reflection;
+4. D stays warm but not camping, reunion, or victory;
+5. the system can move from anchor boards into shot production without rewriting the visual system.
+
+## Source Files Actually Used
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/characters/character-anchor-review-v2.md`
+- `visual-system/characters/character-transparent-slices-v2-from-anchor-review.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+
+Intake:
+
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+
+## Accepted Anchor Inputs
+
+Character:
+
+```text
+visual-system/characters/character-anchor-v2-codex-posture-test.png
+visual-system/characters/character-transparent-slices-v2-from-anchor-preview.png
+visual-system/characters/character-slice-front-v2-from-anchor.png
+visual-system/characters/character-slice-side-v2-from-anchor.png
+visual-system/characters/character-slice-back-v2-from-anchor.png
+visual-system/characters/character-slice-walk-side-v2-from-anchor.png
+visual-system/characters/character-slice-walk-back-v2-from-anchor.png
+```
+
+Style:
+
+```text
+visual-system/style/style-anchor-board-v1.png
+intake/imgs/A.风景主导型蓝图.png
+intake/imgs/B.高位观察型蓝图.png
+visual-system/style/style-anchor-c-repair-v1.png
+intake/imgs/D.夜路篝火型蓝图.png
+```
+
+Do not use the original warm C blueprint as the active C sample:
+
+```text
+intake/imgs/C.泥泞里的星空蓝图.png
+```
+
+It remains only a problem reference.
+
+## Trial Shot Set
+
+| Shot | Segment | Function | Anchor | Why It Is In Trial |
+| --- | --- | --- | --- | --- |
+| `S-01` | Prelude / 0:00-0:16 | World appears before person; cold frosted muddy road, no person. | A | Tests landscape grammar before character appears. |
+| `S-10` | Section B / 0:32-1:05 | Steel canyon / city pressure. | B | Tests city pressure without tech/UI drift. |
+| `S-46` | Final four lines / 3:58-4:16 | Muddy puddle with broken star reflection and boot edge. | C | Tests the soul image and accepted C repair. |
+| `S-47` | Outro / 4:16-4:54 | Transition from muddy starlight to night road fire. | D | Tests whether C can transition into restrained warmth. |
+| `S-51` | Outro / 4:16-4:54 | Final wide: night road, several far fires, tiny traveler. | D | Tests ending restraint, no victory, no gathering. |
+
+## Per-Shot Trial Briefs
+
+### S-01
+
+Purpose:
+
+- prove the MV begins with world before person;
+- establish A as the main landscape-led grammar;
+- test road, frost, mud, cold dawn, and long distance without needing identity preservation.
+
+Must show:
+
+- 16:9 horizontal;
+- cold frosted muddy road;
+- large landscape / large sky / long road;
+- no person;
+- low saturation and road-film realism.
+
+Must avoid:
+
+- travel advertisement;
+- beautiful tourism sunrise;
+- hero backshot;
+- portrait-MV opening;
+- overly polished scenic landscape.
+
+Acceptance test:
+
+> First impression must be road and world, not a person or a poster.
+
+### S-10
+
+Purpose:
+
+- prove B can translate system pressure into physical space;
+- test the traveler as a small coordinate inside steel/concrete scale;
+- avoid AI/tool/UI literalization.
+
+Must show:
+
+- massive overpass, bridge, or steel-concrete urban edge;
+- wet gray ground, puddles, cold city-edge texture;
+- tiny middle-aged traveler at the bottom or edge of the frame;
+- high-angle or distant observational composition;
+- person about 1/10 to 1/6 of frame.
+
+Must preserve character:
+
+- dark long traveler's coat;
+- old backpack;
+- old boots;
+- upright walking posture, not hunched or collapsed.
+
+Must avoid:
+
+- cyber city;
+- UI/code/screens;
+- clean CBD promotion;
+- neon tech aesthetic;
+- centered hero stance.
+
+Acceptance test:
+
+> The city must feel like a canyon pressing the traveler small, not a city promotion image.
+
+### S-46
+
+Purpose:
+
+- prove the accepted C repair can become a shot image;
+- test muddy starlight as the MV's soul image;
+- ensure poetry comes from real mud, not fantasy.
+
+Must show:
+
+- low ground-level composition;
+- irregular muddy puddle as main subject;
+- real wet mud, grit, stones, shoe marks;
+- broken star reflection clearly visible in dark water;
+- old boot / lower-body trace only at the edge, under about 15%;
+- night or deep pre-dawn.
+
+Must preserve character:
+
+- only peripheral identity cues are needed;
+- use boot/lower-body trace, not face;
+- boot must be old and road-worn, not a product closeup.
+
+Must avoid:
+
+- sunrise / sunset;
+- broad warm horizon;
+- magic portal;
+- clean mirror lake;
+- shoe advertisement;
+- only mud/no stars;
+- only stars/no mud.
+
+Acceptance test:
+
+> The viewer first sees real mud, then discovers broken starlight inside it.
+
+### S-47
+
+Purpose:
+
+- test the emotional transition from C to D;
+- turn star reflection into road-fire warmth without becoming camping or reunion;
+- prove D can follow C while staying restrained.
+
+Must show:
+
+- night road after the muddy starlight moment;
+- small local fire near the road;
+- traveler in back or side-back view, not portrait;
+- wet dark road / dark wilderness texture;
+- warm fire only as a small local signal.
+
+Must preserve character:
+
+- same long coat, backpack, and old boots;
+- tired but dignified posture;
+- no collapse, no hunched walk.
+
+Must avoid:
+
+- camping advertisement;
+- large campfire party;
+- group gathering;
+- triumphant rest;
+- cozy lifestyle scene.
+
+Acceptance test:
+
+> The image should say "he can continue", not "he has arrived".
+
+### S-51
+
+Purpose:
+
+- test the final wide ending;
+- prove D's warmth can remain distant and unfinished;
+- ensure the MV closes on road/world, not victory or reunion.
+
+Must show:
+
+- very wide night road or wilderness road;
+- tiny traveler walking away, about 1/20 to 1/12 of frame;
+- one small near fire and one or two distant weak fire points;
+- deep blue-black night with readable dark depth;
+- road and world larger than the person.
+
+Must preserve character:
+
+- recognition mainly through silhouette, long coat, backpack, and walking direction;
+- no face needed;
+- person remains small.
+
+Must avoid:
+
+- visible group reunion;
+- victory image;
+- camping site;
+- large theatrical firelight;
+- inspirational poster backshot;
+- final destination feeling.
+
+Acceptance test:
+
+> It should feel unfinished but less alone: the traveler is still on the road, and far away there are other fires.
+
+## Execution Sequence After This Plan Is Approved
+
+Do not generate all five images at once.
+
+Recommended gated sequence:
+
+1. Create a `gpt-image-2` Skill-based prompt/spec for `S-01`.
+2. User reviews `S-01` prompt/spec.
+3. Generate `S-01` only after approval.
+4. Review `S-01` result and record pass/fail.
+5. Repeat for `S-10`, `S-46`, `S-47`, and `S-51`.
+
+Reason:
+
+- `S-01` is the lowest identity risk and tests world-first composition.
+- `S-10` then tests character-in-environment and city drift.
+- `S-46` tests the highest-risk soul image.
+- `S-47` tests C-to-D transition.
+- `S-51` tests final emotional restraint.
+
+## Proposed File Layout
+
+After approval, create one folder per trial shot only when that shot begins:
+
+```text
+slides/s01/
+slides/s10/
+slides/s46/
+slides/s47/
+slides/s51/
+```
+
+Each shot folder should contain:
+
+```text
+shot-brief-v1.md
+shot-prompt-v1.md
+shot-review-v1.md
+```
+
+Generated images should use stable names:
+
+```text
+shot-image-v1.png
+shot-image-v2.png
+```
+
+If a shot needs revision, revise only that shot. Do not regenerate the whole batch.
+
+## Shared Review Gates
+
+Every generated trial image must pass these before acceptance:
+
+1. 16:9 horizontal.
+2. Matches assigned anchor A/B/C/D.
+3. Does not become portrait, fashion, tourism, city-promo, camping, tech-UI, or victory imagery.
+4. Character, when present, matches the accepted v2 traveler and avoids hunched/collapsed posture.
+5. Environment is larger than the person.
+6. Visual texture remains documentary-real even when poetic.
+7. The image supports later full-shot batching without changing the visual system.
+
+## Trial Success Definition
+
+The five-shot trial succeeds only if:
+
+- `S-01` establishes the road-film world;
+- `S-10` proves city pressure without tech drift;
+- `S-46` proves muddy starlight as the soul image;
+- `S-47` proves restrained fire transition after C;
+- `S-51` proves the ending remains unfinished, quiet, and not victorious.
+
+If two or more shots fail because of the same drift, update the relevant anchor rule before continuing.
+
+If only one shot fails, repair that shot's prompt/spec and retry that shot only.
+
+## Next Gated Action
+
+After this plan is approved, begin with:
+
+```text
+slides/s01/shot-brief-v1.md
+```
+
+Created:
+
+```text
+slides/s01/shot-brief-v1.md
+slides/s01/shot-prompt-v1.md
+slides/s01/shot-image-v1.png
+slides/s01/shot-review-v1.md
+slides/s10/shot-brief-v1.md
+slides/s10/shot-prompt-v1.md
+slides/s10/shot-image-v1.png
+slides/s10/shot-review-v1.md
+slides/s46/shot-brief-v1.md
+slides/s46/shot-prompt-v1.md
+slides/s46/shot-image-v1.png
+slides/s46/shot-review-v1.md
+slides/s47/shot-brief-v1.md
+slides/s47/shot-prompt-v1.md
+slides/s47/shot-image-v1.png
+slides/s47/shot-review-v1.md
+slides/s51/shot-brief-v1.md
+slides/s51/shot-prompt-v1.md
+slides/s51/shot-image-v1.png
+slides/s51/shot-review-v1.md
+```
+
+Current gate:
+
+- five-shot image trial accepted by user;
+- next decision is whether to switch from gated single-shot generation to broad lightweight production.
diff --git a/projects/2026-06-23-在路上/slides/s01/shot-brief-v1.md b/projects/2026-06-23-在路上/slides/s01/shot-brief-v1.md
new file mode 100644
index 0000000..d984e13
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s01/shot-brief-v1.md
@@ -0,0 +1,211 @@
+# S-01 Shot Brief V1
+
+## Status
+
+- Document status: user approved for prompt/spec creation.
+- Execution status: S-01 image accepted by user.
+- Project: `在路上`.
+- Unit: `s01`.
+- Source shot: `S-01`.
+- Segment / time: Prelude / `0:00-0:16`.
+- Trial role: first shot in the five-shot visual test package.
+- Style anchor: A｜风景主导型.
+
+## Purpose
+
+S-01 tests whether the MV can begin with the world before the traveler.
+
+This is the lowest identity-risk trial shot because the traveler should not appear.
+
+It must prove:
+
+- the opening can be landscape-led, not portrait-led;
+- A can establish the main road-film grammar;
+- the road, mud, frost, cold distance, and empty world can carry the emotional premise before any character appears.
+
+Core test sentence:
+
+> 路先于人出现，世界先于人出现。
+
+## Source Files Actually Used
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+
+Intake:
+
+- `intake/阶段五-镜头清单骨架.md`
+
+## Accepted Reference Assets
+
+Primary style reference:
+
+```text
+intake/imgs/A.风景主导型蓝图.png
+```
+
+Style board reference:
+
+```text
+visual-system/style/style-anchor-board-v1.png
+```
+
+Character references:
+
+```text
+none for S-01
+```
+
+Reason:
+
+- S-01 should contain no person.
+- Character references are not needed until shots where the traveler appears.
+- This protects the opening from becoming a portrait-MV or hero-backshot image.
+
+## Shot Definition
+
+Source table definition:
+
+> World appears before person; cold frosted muddy road, no person.
+
+Expanded execution definition:
+
+> A cold frosted muddy road appears in a large empty world before the traveler enters the MV. The viewer should feel the road, distance, damp ground, and cold air before thinking about a person.
+
+## Required Visual Content
+
+The shot must include:
+
+- 16:9 horizontal frame;
+- large landscape / world-first composition;
+- muddy road or worn dirt road extending into distance;
+- cold dawn, pre-dawn, or early morning atmosphere;
+- thin frost or cold wet ground texture;
+- low saturation;
+- documentary-real road-film texture;
+- no visible person.
+
+The road should feel long enough to carry the whole MV's premise:
+
+> 这不是一个人要展示自己，而是一个人即将进入一条比自己更长的路。
+
+## Composition Rules
+
+Required composition:
+
+- environment first;
+- road as primary line;
+- wide or very wide framing;
+- enough sky / ground / distant world to make the frame breathe;
+- no centered hero subject;
+- no foreground object taking over the shot.
+
+Recommended layout:
+
+- road begins near lower frame and travels toward the distance;
+- horizon sits in upper-middle or upper-third area;
+- cold sky and rough ground both remain visible;
+- mud/frost detail is readable but not a closeup.
+
+## Color And Light
+
+Preferred:
+
+- cold gray;
+- earth brown;
+- wet dark blue-gray;
+- subdued frost white;
+- very restrained pale dawn edge if needed.
+
+Avoid:
+
+- golden tourism sunrise;
+- warm inspirational glow;
+- saturated scenic landscape;
+- commercial orange-blue blockbuster grading;
+- bright noon.
+
+Important note:
+
+- The accepted A source image contains a warmer horizon.
+- S-01 may keep a faint cold dawn edge, but must not become a beautiful sunrise postcard.
+
+## Texture Requirements
+
+Must feel real:
+
+- wet mud;
+- rough road ruts;
+- frost or cold dampness;
+- worn ground;
+- natural sky;
+- distant empty terrain.
+
+Do not over-polish the ground or make the scene clean.
+
+The opening should feel quiet and physically cold, not decorative.
+
+## Forbidden Drift
+
+Reject if the image becomes:
+
+- travel advertisement;
+- scenic tourism image;
+- inspirational poster;
+- hero backshot;
+- portrait MV opening;
+- clean fantasy landscape;
+- obvious AI concept-art spectacle;
+- over-bright sunrise;
+- road with a clearly visible person;
+- text, logo, UI, symbols, watermark.
+
+## Relation To Other Trial Shots
+
+S-01 sets the baseline for all later shots:
+
+- before S-10 city pressure, the road/world must already exist;
+- before S-46 muddy starlight, mud and road texture must already be real;
+- before S-47/S-51 firelight, the film must already feel restrained and landscape-led.
+
+If S-01 fails into postcard scenery or hero imagery, the later full MV will likely drift.
+
+## Acceptance Criteria
+
+S-01 passes only if:
+
+1. The first impression is road/world/environment, not person.
+2. No visible traveler appears.
+3. The road feels long, cold, muddy, and real.
+4. The image belongs with the accepted A/B/C/D style board.
+5. The shot avoids tourism, inspirational, and scenic-ad drift.
+6. The frame can plausibly open the MV before the character enters.
+
+## Next Gated Action
+
+This brief has been approved and used to create:
+
+```text
+slides/s01/shot-prompt-v1.md
+```
+
+Generated after prompt approval:
+
+```text
+slides/s01/shot-image-v1.png
+slides/s01/shot-review-v1.md
+```
+
+Current gate:
+
+- `slides/s01/shot-image-v1.png` accepted by user;
+- next trial shot may proceed to `S-10`.
diff --git a/projects/2026-06-23-在路上/slides/s01/shot-image-v1.png b/projects/2026-06-23-在路上/slides/s01/shot-image-v1.png
new file mode 100644
index 0000000..60ae9c0
Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s01/shot-image-v1.png differ
diff --git a/projects/2026-06-23-在路上/slides/s01/shot-prompt-v1.md b/projects/2026-06-23-在路上/slides/s01/shot-prompt-v1.md
new file mode 100644
index 0000000..64448d5
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s01/shot-prompt-v1.md
@@ -0,0 +1,282 @@
+# S-01 Shot Prompt V1
+
+## Status
+
+- Document status: user approved for one S-01 generation.
+- Execution status: used once for Mode B host image generation; candidate image accepted by user.
+- Project: `在路上`.
+- Unit: `s01`.
+- Source shot: `S-01`.
+- Segment / time: Prelude / `0:00-0:16`.
+- Style anchor: A｜风景主导型.
+- Purpose: generate the first trial image for the world-before-person opening.
+- Generated candidate: `shot-image-v1.png`.
+
+## Skill Mode Check
+
+Command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+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)"
+}
+```
+
+Interpretation:
+
+- Use `gpt-image-2` as the prompt-engineering workflow.
+- Garden local generation is not enabled.
+- This Codex environment has a host image-generation path, so later execution should be treated as Mode B if the user approves generation.
+- Do not call Garden scripts for this step.
+
+## Skill Files Actually 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`
+
+Template adaptation:
+
+- `concept-scene.md` is used only for foreground / midground / background / light structure.
+- `minimalist-mood-scene.md` is used for restraint, quiet mood, and avoiding overfilled spectacle.
+- Final output should be a realistic cinematic documentary still, not concept art and not illustration.
+
+## Source Files Actually Used
+
+Primary shot source:
+
+- `slides/s01/shot-brief-v1.md`
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+
+Intake check:
+
+- `intake/阶段五-镜头清单骨架.md`
+
+## Reference Assets
+
+Primary style reference:
+
+```text
+intake/imgs/A.风景主导型蓝图.png
+```
+
+Role:
+
+- A style reference;
+- composition reference;
+- not an edit target;
+- not a requirement to preserve the warm horizon exactly.
+
+Reference dimensions:
+
+```text
+1672x941
+```
+
+Style board reference:
+
+```text
+visual-system/style/style-anchor-board-v1.png
+```
+
+Role:
+
+- common style anchor reference;
+- use for overall palette, road-film grammar, and drift controls.
+
+Character references:
+
+```text
+none
+```
+
+Reason:
+
+- S-01 must contain no visible traveler.
+- Character references should not be attached for this shot because they increase the risk of accidental person insertion.
+
+## S-01 Target
+
+This image must answer:
+
+> 路先于人出现，世界先于人出现。
+
+It is the opening visual premise of the MV.
+
+The viewer should first see:
+
+- a cold world;
+- a long muddy road;
+- frost / wet ground;
+- distance;
+- absence of the traveler.
+
+The image should not yet introduce the character.
+
+## Rendered Prompt
+
+Use case: photorealistic-natural.
+
+Asset type: 16:9 MV trial shot still for `在路上`, `S-01`, opening prelude.
+
+Primary request:
+
+Generate one 16:9 horizontal realistic cinematic documentary still for the opening shot of the `在路上` MV. This is the first image before the traveler appears. The world and the road must appear before any person.
+
+Scene:
+
+A cold frosted muddy road in a large empty wilderness at deep dawn or cold early morning. The road is worn, wet, and rutted, beginning near the lower foreground and extending far into the distance. The ground is heavy with damp mud, rough earth, frozen puddle edges, thin frost, small stones, dirty road texture, and cold moisture. The surrounding landscape is open, quiet, and sparse, with distant low hills or empty terrain fading into cold haze.
+
+Composition:
+
+Wide or very wide road-film composition. The frame should be dominated by the road, ground, sky, and distance. The road is the main visual line. The horizon sits in the upper-middle or upper-third area. Leave enough sky and empty terrain to make the world feel larger than any future person. No foreground object should become the main subject. No visible traveler, no human silhouette, no back view, no vehicle, no animal.
+
+Style:
+
+Real documentary texture + restrained poetic realism + landscape-led road-film composition. It should feel physically cold and quiet, not decorative. The image should belong with a low-saturation MV style anchor board: cold gray, earth brown, dark khaki, coal black, wet dark blue-gray, subdued frost white.
+
+Lighting:
+
+Cold pre-dawn or early morning natural light. Very restrained pale silver-blue dawn edge is allowed. No visible sun. No broad golden sunrise. No warm inspirational glow. No bright noon. No theatrical rim light. Dark and midtone areas should retain real depth and texture.
+
+Mood:
+
+Quiet, empty, cold, long-road feeling. The emotional center is not victory and not despair. It is the beginning of continuation: the road is already there before the person enters.
+
+Strict constraints:
+
+- 16:9 horizontal frame;
+- no visible person;
+- no traveler silhouette;
+- no hero backshot;
+- no portrait subject;
+- road and world first;
+- muddy road and frost must feel physically real;
+- low saturation;
+- documentary realism;
+- no text, no labels, no logo, no UI, no watermark.
+
+Strictly avoid:
+
+- travel advertisement;
+- scenic tourism landscape;
+- beautiful postcard sunrise;
+- golden sunset or golden sunrise;
+- motivational poster;
+- fantasy landscape;
+- epic concept-art spectacle;
+- clean polished road;
+- fashion/travel-blog image;
+- visible person or human silhouette;
+- vehicle tracks as the main subject;
+- buildings or city skyline;
+- camping/firelight;
+- neon/cyber/sci-fi;
+- overly saturated colors;
+- text, symbols, watermarks, UI.
+
+Final image behavior:
+
+The viewer should first read a cold muddy road and a large empty world. The absence of the traveler should be clear. This frame should plausibly open the MV before the character enters.
+
+## Reference Handling At Generation Time
+
+Current recommended execution:
+
+- Use the rendered prompt as a standalone prompt.
+
+If the host image tool supports image references:
+
+- attach `intake/imgs/A.风景主导型蓝图.png` as style / composition reference;
+- optionally attach `visual-system/style/style-anchor-board-v1.png` as global style reference;
+- explicitly instruct the model not to copy the warm horizon too strongly;
+- do not attach character references.
+
+If the host image tool does not support image references:
+
+- use the rendered prompt as a standalone prompt;
+- preserve the same A-anchor analysis in the review file.
+
+## Generation Result
+
+Generated after user approval:
+
+```text
+slides/s01/shot-image-v1.png
+```
+
+Review file:
+
+```text
+slides/s01/shot-review-v1.md
+```
+
+Generation mode:
+
+```text
+gpt-image-2 Skill Mode B; host image-generation tool; no Garden script.
+```
+
+## Acceptance Criteria
+
+S-01 passes only if:
+
+1. The first impression is road/world/environment, not person.
+2. No visible traveler, human silhouette, vehicle, or animal appears.
+3. The road feels long, cold, muddy, and real.
+4. The image belongs with the accepted common style anchor board.
+5. The shot avoids tourism, inspirational, scenic-ad, fantasy, and hero-backshot drift.
+6. The frame can plausibly open the MV before the character enters.
+
+## Failure Conditions
+
+Reject the generation if any of these happen:
+
+- a person appears;
+- a distant silhouette reads as the traveler;
+- the image becomes a beautiful sunrise postcard;
+- the image becomes an inspirational road poster;
+- the road or mud looks too clean;
+- the scene feels like travel advertising;
+- colors become too saturated or broadly golden;
+- the frame no longer belongs with A/B/C/D.
+
+## Review File Created After Generation
+
+After user approval and generation, created:
+
+```text
+slides/s01/shot-review-v1.md
+```
+
+That review compares the generated image against:
+
+- `slides/s01/shot-brief-v1.md`;
+- this prompt/spec;
+- accepted A style reference;
+- accepted common style anchor board;
+- the S-01 acceptance criteria above.
diff --git a/projects/2026-06-23-在路上/slides/s01/shot-review-v1.md b/projects/2026-06-23-在路上/slides/s01/shot-review-v1.md
new file mode 100644
index 0000000..ba0f4bb
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s01/shot-review-v1.md
@@ -0,0 +1,157 @@
+# S-01 Shot Review V1
+
+## Status
+
+- Document status: user accepted.
+- Execution status: S-01 image generated once; no other trial shots generated.
+- Project: `在路上`.
+- Unit: `s01`.
+- Source shot: `S-01`.
+- Segment / time: Prelude / `0:00-0:16`.
+- Style anchor: A｜风景主导型.
+- Candidate image: `shot-image-v1.png`.
+- Review result: accepted by user.
+
+## Generation Trace
+
+Skill mode check command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+Skill mode check 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 · 未启用 Garden。若宿主 Agent 自带图像工具（image_generation / dalle / mcp__*image* 等）→ MODE B：把 prompt 交给宿主出图。若宿主无图像工具 → MODE C：仅产出高质量 prompt 给用户。"
+}
+```
+
+Interpretation:
+
+- Garden local generation was not enabled.
+- The host image-generation path was available.
+- This generation used `gpt-image-2` Skill Mode B: render prompt through the Skill workflow, then pass the approved prompt to the host image tool.
+- No Garden generation script was called.
+
+Prompt/spec used:
+
+```text
+slides/s01/shot-prompt-v1.md
+```
+
+Host default output:
+
+```text
+C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0e239cf2ce5b548c016a3a4ec0ad44819187378af2c5b94dd6.png
+```
+
+Project copy:
+
+```text
+slides/s01/shot-image-v1.png
+```
+
+The original host output was left in place.
+
+## Image File Check
+
+Project image path:
+
+```text
+C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\slides\s01\shot-image-v1.png
+```
+
+File size:
+
+```text
+2432836 bytes
+```
+
+Dimensions:
+
+```text
+1672x941
+```
+
+Pixel format:
+
+```text
+Format24bppRgb
+```
+
+Aspect note:
+
+- The image is horizontal and close to 16:9.
+- It is not mathematically exact 16:9 because `1672 / 941 = 1.7768`, while exact 16:9 is `1.7777`.
+- No crop or resize was applied in this review step.
+
+## Source Alignment
+
+Primary sources checked:
+
+- `slides/s01/shot-brief-v1.md`
+- `slides/s01/shot-prompt-v1.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Reference asset handling:
+
+- No character reference was used.
+- No person was intentionally introduced.
+- The prompt used the A-anchor road/world grammar as text guidance.
+- The host image tool was used as a standalone prompt path; image references were not attached in this run.
+
+## Local Inspection
+
+Local visual inspection:
+
+- The first read is a cold muddy road and a large empty world.
+- No visible traveler, human silhouette, vehicle, or animal is apparent.
+- The road is wet, rutted, and physically textured.
+- Frost/cold dampness is readable on the surrounding ground.
+- The palette stays low-saturation: blue-gray sky, dark wet mud, subdued earth tones.
+- The composition is landscape-led and wide, with the road pulling into distant terrain.
+- The frame does not read as tourism, fashion, city-promo, camping, fantasy, or hero-backshot imagery.
+
+Watch points:
+
+- The horizon contains a pale dawn band. It stays restrained in this candidate, but should be checked by user review so it does not drift toward postcard sunrise in later variants.
+- The image is very close to 16:9 but not exact.
+
+External vision analysis:
+
+- Not used for this review.
+- The attempted external image-understanding path was rejected because it would transmit a private project image to an unapproved external service.
+- This review therefore relies on local visual inspection and local file metadata only.
+
+## Acceptance Criteria Review
+
+| Criterion | Assessment | Notes |
+| --- | --- | --- |
+| First impression is road/world/environment, not person | pass for review | Road, mud, sky, and distance dominate the frame. |
+| No visible traveler, human silhouette, vehicle, or animal | pass for review | No visible figure or moving object found in local inspection. |
+| Road feels long, cold, muddy, and real | pass for review | Wet ruts, puddles, stones, and frost/cold ground are readable. |
+| Belongs with accepted common style anchor board | pass for review | Low-saturation road-film palette and restrained landscape grammar match A. |
+| Avoids tourism, inspirational, scenic-ad, fantasy, hero-backshot drift | pass for review | No hero subject, no saturated sunrise, no fantasy spectacle. |
+| Plausibly opens the MV before the character enters | pass for review | It establishes the road and world before any person appears. |
+
+## Decision Needed
+
+User accepted:
+
+```text
+slides/s01/shot-image-v1.png
+```
+
+Next gated step can begin for `S-10`.
+
+S-01 is closed unless a later consistency pass requires revision.
diff --git a/projects/2026-06-23-在路上/slides/s10/shot-brief-v1.md b/projects/2026-06-23-在路上/slides/s10/shot-brief-v1.md
new file mode 100644
index 0000000..4ae4c3d
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s10/shot-brief-v1.md
@@ -0,0 +1,199 @@
+# S-10 Shot Brief V1
+
+## Status
+
+- Document status: created for direct S-10 generation.
+- Execution status: S-10 image accepted by user.
+- Project: `在路上`.
+- Unit: `s10`.
+- Source shot: `S-10`.
+- Segment / time: Section B / `0:32-1:05`.
+- Trial role: second shot in the five-shot visual test package.
+- Style anchor: B｜高位观察型 / 城市边缘纪录片感.
+
+## Purpose
+
+S-10 tests whether the accepted visual system can translate pressure from "system" into physical space without showing technology literally.
+
+Core test sentence:
+
+> 城市变成峡谷，人物变成底部小坐标。
+
+This shot should not be a city promotion image. It should make the traveler feel small inside a cold steel / concrete structure.
+
+## Source Files Actually Used
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/characters/character-anchor-review-v2.md`
+- `visual-system/characters/character-transparent-slices-v2-from-anchor-review.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+
+Intake:
+
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+
+## Accepted Reference Assets
+
+Primary style reference:
+
+```text
+intake/imgs/B.高位观察型蓝图.png
+```
+
+Style board reference:
+
+```text
+visual-system/style/style-anchor-board-v1.png
+```
+
+Character references:
+
+```text
+visual-system/characters/character-anchor-v2-codex-posture-test.png
+visual-system/characters/character-transparent-slices-v2-from-anchor-preview.png
+visual-system/characters/character-slice-walk-side-v2-from-anchor.png
+visual-system/characters/character-slice-walk-back-v2-from-anchor.png
+```
+
+Role:
+
+- B image: style / composition reference for high observation, cold city edge, wet ground, and small person.
+- Style board: global palette and drift-control reference.
+- Character assets: identity, clothing, backpack, boots, and upright walking posture reference.
+
+## Shot Definition
+
+Source table definition:
+
+> Steel canyon / city pressure.
+
+Expanded execution definition:
+
+> A tiny middle-aged traveler walks through the bottom of a massive steel-concrete urban edge. Bridge columns, overpass shadows, cold walls, wet ground, and structural scale create a canyon feeling. The shot should make modern systems feel heavy without showing UI, screens, code, or futuristic technology.
+
+## Required Visual Content
+
+The shot must include:
+
+- 16:9 horizontal frame;
+- high-angle or distant observational composition;
+- urban edge, massive overpass, bridge columns, or steel-concrete canyon;
+- wet gray ground, puddles, cold concrete, steel beams, and small rust traces;
+- tiny traveler at the bottom or edge of frame, about `1/10` to `1/6` of the frame;
+- long dark traveler's coat, old backpack, old boots;
+- upright walking posture, not hunched or collapsed;
+- low-saturation documentary realism.
+
+## Composition Rules
+
+Required composition:
+
+- the structure is larger than the person;
+- the person is a moving coordinate, not the main hero;
+- overhead pressure, side pressure, and forward length should all be visible;
+- no centered hero pose;
+- no face-first or portrait composition.
+
+Recommended layout:
+
+- overhead or high-observation angle looking down into a concrete / steel channel;
+- person placed at lower third, lower edge, or side-bottom path;
+- wet ground and vertical supports create geometric pressure.
+
+## Character Rules
+
+The traveler should be recognizable mainly through:
+
+- black rectangular glasses only if visible, but face detail is not required;
+- longer messy dark hair outline;
+- longer dense beard / stubble only if the side view is readable;
+- dark gray / dark khaki / coal-black long road-worn coat;
+- old medium-volume backpack;
+- old muddy boots;
+- upright chest and straight walking spine.
+
+Do not make him:
+
+- hunched;
+- collapsed;
+- homeless-coded;
+- drunk-coded;
+- a fashion model;
+- a heroic poster figure;
+- a city explorer / travel blogger.
+
+## Color And Light
+
+Preferred:
+
+- steel gray;
+- concrete gray;
+- wet blue-gray;
+- coal black;
+- dark khaki;
+- small rust-brown traces.
+
+Lighting:
+
+- overcast;
+- bridge shadow;
+- cold reflected light from wet ground;
+- no bright sun;
+- no dramatic commercial rim light.
+
+Avoid:
+
+- neon blue / purple;
+- cyberpunk glow;
+- orange-blue blockbuster city grading;
+- polished commercial skyline.
+
+## Forbidden Drift
+
+Reject if the image becomes:
+
+- futuristic city;
+- AI interface / code / screen / UI image;
+- clean CBD promotion;
+- glass-tower city advertisement;
+- cyberpunk / neon / sci-fi;
+- fashion street photo;
+- heroic centered pose;
+- portrait MV image;
+- the traveler is bent, collapsed, or spiritually defeated;
+- text, logo, UI, watermark, readable signage as main subject.
+
+## Acceptance Criteria
+
+S-10 passes only if:
+
+1. The city edge feels oppressive.
+2. The traveler is small enough.
+3. The frame has a steel / concrete canyon feeling.
+4. It reads as city edge, not prosperous city promotion.
+5. There is no tech UI, code, screen, or sci-fi drift.
+6. The traveler is walking naturally, not posing.
+7. It can connect from "gravity downward" into "where is home".
+
+## Next Action
+
+Created and used for direct prompt/spec and image generation:
+
+```text
+slides/s10/shot-prompt-v1.md
+slides/s10/shot-image-v1.png
+slides/s10/shot-review-v1.md
+```
+
+No separate prompt review was required for this run because the user explicitly requested direct S-10 image generation and final image review only.
diff --git a/projects/2026-06-23-在路上/slides/s10/shot-image-v1.png b/projects/2026-06-23-在路上/slides/s10/shot-image-v1.png
new file mode 100644
index 0000000..5151e48
Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s10/shot-image-v1.png differ
diff --git a/projects/2026-06-23-在路上/slides/s10/shot-prompt-v1.md b/projects/2026-06-23-在路上/slides/s10/shot-prompt-v1.md
new file mode 100644
index 0000000..a5b17eb
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s10/shot-prompt-v1.md
@@ -0,0 +1,305 @@
+# S-10 Shot Prompt V1
+
+## Status
+
+- Document status: created for direct generation.
+- Execution status: used once for Mode B host image generation; candidate image accepted by user.
+- Project: `在路上`.
+- Unit: `s10`.
+- Source shot: `S-10`.
+- Segment / time: Section B / `0:32-1:05`.
+- Style anchor: B｜高位观察型 / 城市边缘纪录片感.
+- Purpose: generate the second trial image for the steel canyon / city pressure shot.
+- Generated candidate: `shot-image-v1.png`.
+
+## User Gate
+
+User accepted S-01 and changed the trial workflow:
+
+```text
+S-01 passed. For S-10, skip intermediate prompt review and generate the image directly. User will review the final image.
+```
+
+This file still records the prompt/spec first because all project-bound image prompts in this repository must be produced through the `gpt-image-2` Skill workflow.
+
+## Skill Mode Check
+
+Command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+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 · 未启用 Garden。若宿主 Agent 自带图像工具（image_generation / dalle / mcp__*image* 等）→ MODE B：把 prompt 交给宿主出图。若宿主无图像工具 → MODE C：仅产出高质量 prompt 给用户。"
+}
+```
+
+Interpretation:
+
+- Use `gpt-image-2` as the prompt-engineering workflow.
+- Garden local generation is not enabled.
+- This Codex environment has a host image-generation path, so execution is Mode B.
+- Do not call Garden scripts for this step.
+
+## Skill Files Actually 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\.agents\skills\gpt-image-2\references\portraits-and-characters\pose-reference-sheet.md`
+
+Template adaptation:
+
+- `concept-scene.md` is used for foreground / midground / background / scale-pressure structure.
+- `minimalist-mood-scene.md` is used for restraint and avoiding overfilled city spectacle.
+- `pose-reference-sheet.md` is used only for fixed character-attribute discipline; this is not a pose sheet.
+- Final output should be a realistic cinematic documentary still, not concept art, not illustration, and not a character poster.
+
+## Source Files Actually Used
+
+Primary shot source:
+
+- `slides/s10/shot-brief-v1.md`
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/characters/character-anchor-review-v2.md`
+- `visual-system/characters/character-transparent-slices-v2-from-anchor-review.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+
+Intake:
+
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+
+## Reference Assets
+
+Primary style reference:
+
+```text
+intake/imgs/B.高位观察型蓝图.png
+```
+
+Role:
+
+- B style reference;
+- high-observation / city-edge reference;
+- not an edit target.
+
+Style board reference:
+
+```text
+visual-system/style/style-anchor-board-v1.png
+```
+
+Role:
+
+- common style anchor reference;
+- use for palette, documentary restraint, and drift controls.
+
+Character references:
+
+```text
+visual-system/characters/character-anchor-v2-codex-posture-test.png
+visual-system/characters/character-transparent-slices-v2-from-anchor-preview.png
+visual-system/characters/character-slice-walk-side-v2-from-anchor.png
+visual-system/characters/character-slice-walk-back-v2-from-anchor.png
+```
+
+Role:
+
+- identity / clothing / prop / posture references;
+- not edit targets.
+
+## S-10 Target
+
+This image must answer:
+
+> 一个中年行者，被巨大的现代系统压到很小。
+
+The shot must translate pressure into space:
+
+- steel and concrete are large;
+- the traveler is small;
+- the city is edge-like and cold;
+- the system is implied through structure, not shown as UI or technology.
+
+## Rendered Prompt
+
+Use case: photorealistic-natural.
+
+Asset type: 16:9 MV trial shot still for `在路上`, `S-10`, Section B city-pressure shot.
+
+Primary request:
+
+Generate one 16:9 horizontal realistic cinematic documentary still for the `S-10` steel canyon shot of the `在路上` MV. The image should show a tiny middle-aged traveler walking through the bottom of a massive steel-concrete urban edge, as if the modern city structure has become a canyon pressing him small. This is not a city promotion image and not a technology image.
+
+Scene:
+
+An urban edge under a massive overpass, bridge structure, or steel-concrete canyon on a cold overcast day. Huge bridge columns, concrete walls, steel beams, damp retaining walls, wet gray ground, puddles, drain lines, worn road texture, small rust stains, and distant city walls or building gaps form a heavy structural space. The place feels like the outside edge of a city, not its bright center. It is wet, cold, gray, and physically real.
+
+Subject:
+
+A tiny 50-year-old male traveler, about 180cm tall, walking through the bottom or side-bottom of the frame. He wears a dark gray / dark khaki / coal-black long road-worn coat, dark inner layer, old dark pants, old muddy walking boots, and an old practical backpack. He has longer slightly messy dark hair and a dense controlled beard / stubble. His face is not the focus; recognition should come mainly from the upright walking silhouette, long coat, backpack, and boots.
+
+Character posture:
+
+The traveler is walking naturally and quietly, not posing. His walking posture must be upright with a straighter spine, open chest, natural head and neck alignment, and relaxed shoulders. He may look tired, but he must not be hunched, bent at the waist, rounded forward, bowed in defeat, collapsed, frail, drunk, homeless-coded, or spiritually broken.
+
+Composition:
+
+High-angle or distant observational road-film composition. The structure must be larger than the person. The traveler should occupy roughly 1/10 to 1/6 of the frame, never becoming a portrait subject. The overhead bridge mass presses down, the two sides close in, and the forward path runs long through the canyon. Place the traveler near the lower third, lower edge, or side-bottom path. Use concrete geometry, vertical supports, wet ground, and puddles to create pressure. No centered hero stance.
+
+Style:
+
+Real documentary texture + restrained poetic realism + city-edge road-film composition. Low saturation. The image should belong with the accepted `在路上` style anchor board: cold gray, concrete gray, steel gray, wet dark blue-gray, dark khaki, coal black, small rust-brown traces. It should feel observed and physical, not decorative.
+
+Lighting:
+
+Cold overcast natural light, bridge shadow, wet ground reflection, low contrast, no bright sun. No theatrical rim light. No neon, no sci-fi glow, no cyberpunk blue-purple, no orange-blue blockbuster city grading. Dark and midtone areas should retain real concrete and wet-ground texture.
+
+Mood:
+
+Quiet pressure, system weight, city gravity, no home. The emotional center is not action and not defeat. It is a small person continuing to walk through a structure much larger than himself.
+
+Strict constraints:
+
+- 16:9 horizontal frame;
+- high-angle or distant observational shot;
+- massive overpass / bridge / steel-concrete canyon;
+- tiny traveler about 1/10 to 1/6 of frame;
+- traveler is walking, not posing;
+- upright walking posture, not hunched;
+- environment first, person second;
+- wet ground, puddles, cold gray light;
+- documentary realism;
+- low saturation;
+- no text, no labels, no logo, no UI, no watermark.
+
+Strictly avoid:
+
+- futuristic UI;
+- code screens;
+- AI tool interface;
+- sci-fi city;
+- cyberpunk city;
+- neon lighting;
+- clean CBD promotion;
+- glass-tower commercial skyline as the main subject;
+- polished city advertisement;
+- fashion street photography;
+- travel-blog image;
+- heroic centered pose;
+- portrait MV image;
+- direct front-facing portrait;
+- arms-open inspirational posture;
+- hunched back;
+- bent waist;
+- collapsed defeated body;
+- homeless or drunk stereotype;
+- bright noon;
+- golden sunset or sunrise;
+- readable signs, logos, text, symbols, watermarks, UI.
+
+Final image behavior:
+
+The viewer should first read a cold steel-concrete canyon and a large oppressive city-edge structure. Then they should notice the traveler as a small moving coordinate at the bottom. The image should plausibly follow `S-09` gravity downward and lead into `S-11` no home / outside the city interior.
+
+## Reference Handling At Generation Time
+
+Current execution:
+
+- Use the rendered prompt as a standalone prompt.
+- The local reference assets were read / viewed for prompt construction.
+- The host image tool path does not provide explicit file-reference arguments in this run.
+
+If a later host path supports image references:
+
+- attach `intake/imgs/B.高位观察型蓝图.png` as style / composition reference;
+- attach `visual-system/style/style-anchor-board-v1.png` as global style reference;
+- attach `visual-system/characters/character-slice-walk-side-v2-from-anchor.png` or `character-slice-walk-back-v2-from-anchor.png` as identity / posture reference;
+- explicitly instruct the model to keep the traveler small and not turn him into a portrait subject.
+
+## Generation Result
+
+Generated after user authorization to skip intermediate prompt review:
+
+```text
+slides/s10/shot-image-v1.png
+```
+
+Review file:
+
+```text
+slides/s10/shot-review-v1.md
+```
+
+Generation mode:
+
+```text
+gpt-image-2 Skill Mode B; host image-generation tool; no Garden script.
+```
+
+## Acceptance Criteria
+
+S-10 passes only if:
+
+1. The city edge feels oppressive.
+2. The traveler is small enough.
+3. The frame has a steel / concrete canyon feeling.
+4. It reads as city edge, not prosperous city promotion.
+5. There is no tech UI, code, screen, or sci-fi drift.
+6. The traveler is walking naturally, not posing.
+7. The traveler is not hunched or collapsed.
+8. It can connect from "gravity downward" into "where is home".
+
+## Failure Conditions
+
+Reject the generation if any of these happen:
+
+- the city becomes a clean CBD advertisement;
+- the image becomes cyberpunk, neon, or sci-fi;
+- text, UI, code, screens, or AI-tool motifs appear;
+- the traveler becomes too large or portrait-like;
+- the traveler stands centered like a hero;
+- the traveler is hunched, collapsed, or bowed in defeat;
+- the scene feels like street fashion or travel photography;
+- the frame lacks steel / concrete canyon pressure.
+
+## Review File Created After Generation
+
+After generation, created:
+
+```text
+slides/s10/shot-review-v1.md
+```
+
+That review compares the generated image against:
+
+- `slides/s10/shot-brief-v1.md`;
+- this prompt/spec;
+- accepted B style reference;
+- accepted common style anchor board;
+- accepted v2 character identity lock;
+- the S-10 acceptance criteria above.
diff --git a/projects/2026-06-23-在路上/slides/s10/shot-review-v1.md b/projects/2026-06-23-在路上/slides/s10/shot-review-v1.md
new file mode 100644
index 0000000..12f09c9
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s10/shot-review-v1.md
@@ -0,0 +1,161 @@
+# S-10 Shot Review V1
+
+## Status
+
+- Document status: user accepted.
+- Execution status: S-10 image generated once.
+- Project: `在路上`.
+- Unit: `s10`.
+- Source shot: `S-10`.
+- Segment / time: Section B / `0:32-1:05`.
+- Style anchor: B｜高位观察型 / 城市边缘纪录片感.
+- Candidate image: `shot-image-v1.png`.
+- Review result: accepted by user.
+
+## Generation Trace
+
+Skill mode check command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+Skill mode check 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 · 未启用 Garden。若宿主 Agent 自带图像工具（image_generation / dalle / mcp__*image* 等）→ MODE B：把 prompt 交给宿主出图。若宿主无图像工具 → MODE C：仅产出高质量 prompt 给用户。"
+}
+```
+
+Interpretation:
+
+- Garden local generation was not enabled.
+- The host image-generation path was available.
+- This generation used `gpt-image-2` Skill Mode B: render prompt through the Skill workflow, then pass the prompt to the host image tool.
+- No Garden generation script was called.
+
+Prompt/spec used:
+
+```text
+slides/s10/shot-prompt-v1.md
+```
+
+Host default output:
+
+```text
+C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0971672cba509111016a3a519ffb908190922c08bb53c61e89.png
+```
+
+Project copy:
+
+```text
+slides/s10/shot-image-v1.png
+```
+
+The original host output was left in place.
+
+## Image File Check
+
+Project image path:
+
+```text
+C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\slides\s10\shot-image-v1.png
+```
+
+File size:
+
+```text
+2235423 bytes
+```
+
+Dimensions:
+
+```text
+1672x941
+```
+
+Pixel format:
+
+```text
+Format24bppRgb
+```
+
+Aspect note:
+
+- The image is horizontal and close to 16:9.
+- It is not mathematically exact 16:9 because `1672 / 941 = 1.7768`, while exact 16:9 is `1.7777`.
+- No crop or resize was applied in this review step.
+
+## Source Alignment
+
+Primary sources checked:
+
+- `slides/s10/shot-brief-v1.md`
+- `slides/s10/shot-prompt-v1.md`
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Reference asset handling:
+
+- Character and style references were read / viewed locally for prompt construction.
+- The host image tool was used as a standalone prompt path; explicit image-reference attachment was not available in this run.
+
+## Local Inspection
+
+Local visual inspection:
+
+- The first read is a dark steel-concrete overpass / bridge structure.
+- The scene reads as a cold city edge rather than a clean city promotion.
+- The traveler is visible near the lower-left area and remains small relative to the structure.
+- The wet ground, puddles, bridge shadow, concrete wall, and overhead mass create pressure.
+- The palette stays low-saturation: concrete gray, steel gray, wet dark blue-gray, coal-black shadows.
+- No obvious futuristic UI, code screen, AI interface, neon, cyberpunk, or sci-fi city motif appears.
+- The traveler does not read as a centered hero, fashion model, or travel blogger.
+
+Watch points:
+
+- The distant skyline is visible, but it is secondary and not glossy enough to read as city promotion.
+- The traveler is small, so identity fidelity is mainly silhouette / coat / backpack / walking scale, not face detail.
+- The walking posture appears upright at this scale, but the small figure means fine posture judgment remains user-review dependent.
+
+External vision analysis:
+
+- Not used for this review.
+- This review relies on local visual inspection and local file metadata only.
+
+## Acceptance Criteria Review
+
+| Criterion | Assessment | Notes |
+| --- | --- | --- |
+| City edge feels oppressive | pass for review | Bridge ceiling, concrete wall, and wet dark ground press the frame. |
+| Traveler is small enough | pass for review | Traveler is a small lower-left coordinate; structure dominates. |
+| Steel / concrete canyon feeling | pass for review | Overpass underside and side wall form a channel-like space. |
+| City edge, not prosperous city promotion | pass for review | Scene is wet, cold, dark, and utilitarian. |
+| No tech UI, code, screen, or sci-fi drift | pass for review | No obvious UI/code/neon/futuristic elements seen. |
+| Traveler walking naturally, not posing | pass for review | Figure appears to be moving through the space rather than posing. |
+| Traveler not hunched or collapsed | mostly pass for review | At this scale the silhouette reads upright; fine identity/posture check remains for user review. |
+| Connects from "gravity downward" into "where is home" | pass for review | The image implies moving through pressure outside the city interior. |
+
+## Decision Needed
+
+User accepted:
+
+```text
+slides/s10/shot-image-v1.png
+```
+
+Next gated shot can begin from `S-46`.
+
+S-10 is closed unless a later consistency pass requires revision.
diff --git a/projects/2026-06-23-在路上/slides/s46/shot-brief-v1.md b/projects/2026-06-23-在路上/slides/s46/shot-brief-v1.md
new file mode 100644
index 0000000..465aa78
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s46/shot-brief-v1.md
@@ -0,0 +1,193 @@
+# S-46 Shot Brief V1
+
+## Status
+
+- Document status: created for direct S-46 generation.
+- Execution status: prompt/spec created and used directly; S-46 image accepted by user.
+- Project: `在路上`.
+- Unit: `s46`.
+- Source shot: `S-46`.
+- Segment / time: Final four lines / `3:58-4:16`.
+- Trial role: third shot in the five-shot visual test package.
+- Style anchor: C｜泥泞里的星空.
+
+## Purpose
+
+S-46 tests whether the accepted C repair can become the MV's soul image as an actual shot.
+
+Core test sentence:
+
+> 低头看泥水，发现星空也在里面。
+
+This shot is not about escaping the mud. It is about meaning appearing inside the mud.
+
+## Source Files Actually Used
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+- `visual-system/style/style-anchor-c-repair-review-v1.md`
+
+Intake:
+
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+
+## Accepted Reference Assets
+
+Primary C reference:
+
+```text
+visual-system/style/style-anchor-c-repair-v1.png
+```
+
+Style board reference:
+
+```text
+visual-system/style/style-anchor-board-v1.png
+```
+
+Character / boot references:
+
+```text
+visual-system/characters/character-anchor-v2-codex-posture-test.png
+visual-system/characters/character-slice-walk-side-v2-from-anchor.png
+```
+
+Role:
+
+- C repair: primary mood / composition / acceptance reference.
+- Style board: global palette and drift-control reference.
+- Character references: only boot / coat-edge / lower-body identity cues; not a portrait or full-body target.
+
+## Shot Definition
+
+Source table definition:
+
+> Muddy puddle with broken star reflection and boot edge.
+
+Expanded execution definition:
+
+> A close-to-ground night or deep pre-dawn frame shows a real muddy puddle as the main subject. Broken star points are reflected in dark water and interrupted by mud, stones, ripples, and shoe marks. Only a small old boot edge, coat edge, or lower-body shadow may appear at the frame edge to connect the image to the traveler.
+
+## Required Visual Content
+
+The shot must include:
+
+- 16:9 horizontal frame;
+- low ground-level composition;
+- irregular muddy puddle as the main subject;
+- real wet mud, grit, stones, shoe marks, and rough puddle edges;
+- dark water with broken star reflection clearly visible;
+- old worn boot / lower-body trace only at the edge, under about `15%`;
+- night or deep pre-dawn darkness;
+- low-saturation poetic documentary realism.
+
+## Composition Rules
+
+Required composition:
+
+- mud and puddle first;
+- star reflection discovered inside the water;
+- boot / person only peripheral;
+- no face, no full body, no portrait;
+- no clean mirror lake;
+- no real sky as the main star field.
+
+Recommended layout:
+
+- puddle occupies center or lower-center;
+- surrounding mud frames the puddle;
+- star reflection appears inside dark water, broken by water ripples, stones, and irregular mud;
+- old boot edge or coat hem enters from a far edge only.
+
+## Character / Boot Rules
+
+Only peripheral identity cues are needed:
+
+- old road-worn boot;
+- muddy boot edge;
+- small coat hem / lower leg shadow if useful.
+
+The boot must not become:
+
+- a product closeup;
+- shoe advertisement;
+- fashion outdoor gear;
+- clean new boot;
+- the main subject.
+
+## Color And Light
+
+Preferred:
+
+- dark blue-gray;
+- wet black-gray;
+- mud brown;
+- coal black;
+- small cold-white star points.
+
+Lighting:
+
+- night or deep pre-dawn darkness;
+- weak water reflection;
+- cold star glints in the puddle;
+- no visible sun;
+- no warm horizon;
+- no firelight interference.
+
+Avoid:
+
+- sunrise or sunset;
+- broad warm light;
+- golden glow;
+- fantasy portal light;
+- bright clean Milky Way spectacle.
+
+## Forbidden Drift
+
+Reject if the image becomes:
+
+- sunrise / sunset puddle;
+- pretty landscape reflection;
+- magic portal;
+- fantasy lake;
+- shoe or boot advertisement;
+- clean mirror water;
+- only mud with no stars;
+- only stars with no mud;
+- full person / face / portrait;
+- text, logo, UI, watermark.
+
+## Acceptance Criteria
+
+S-46 passes only if:
+
+1. The viewer first sees real mud and water.
+2. Broken star reflection is clear enough to read.
+3. The scene is night or deep pre-dawn, not sunrise or sunset.
+4. The boot / traveler trace stays peripheral and under about `15%`.
+5. The boot does not become a product shot.
+6. The image is restrained and documentary-real, not fantasy.
+7. It truly conveys "泥泞里，有星空".
+
+## Next Action
+
+Created and used for direct prompt/spec and image generation:
+
+```text
+slides/s46/shot-prompt-v1.md
+slides/s46/shot-image-v1.png
+slides/s46/shot-review-v1.md
+```
+
+No separate prompt review was required for this run because the user requested moving directly to the next generated shot and final image review.
diff --git a/projects/2026-06-23-在路上/slides/s46/shot-image-v1.png b/projects/2026-06-23-在路上/slides/s46/shot-image-v1.png
new file mode 100644
index 0000000..f1d12cb
Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s46/shot-image-v1.png differ
diff --git a/projects/2026-06-23-在路上/slides/s46/shot-prompt-v1.md b/projects/2026-06-23-在路上/slides/s46/shot-prompt-v1.md
new file mode 100644
index 0000000..f8462fa
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s46/shot-prompt-v1.md
@@ -0,0 +1,301 @@
+# S-46 Shot Prompt V1
+
+## Status
+
+- Document status: created for direct generation.
+- Execution status: used once for Mode B host image generation; candidate image accepted by user.
+- Project: `在路上`.
+- Unit: `s46`.
+- Source shot: `S-46`.
+- Segment / time: Final four lines / `3:58-4:16`.
+- Style anchor: C｜泥泞里的星空.
+- Purpose: generate the third trial image for the muddy starlight soul shot.
+- Generated candidate: `shot-image-v1.png`.
+
+## User Gate
+
+User accepted S-10 and requested the next shot.
+
+The current trial workflow skips intermediate prompt review. The user reviews final generated images.
+
+## Skill Mode Check
+
+Command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+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 · 未启用 Garden。若宿主 Agent 自带图像工具（image_generation / dalle / mcp__*image* 等）→ MODE B：把 prompt 交给宿主出图。若宿主无图像工具 → MODE C：仅产出高质量 prompt 给用户。"
+}
+```
+
+Interpretation:
+
+- Use `gpt-image-2` as the prompt-engineering workflow.
+- Garden local generation is not enabled.
+- This Codex environment has a host image-generation path, so execution is Mode B.
+- Do not call Garden scripts for this step.
+
+## Skill Files Actually 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\.agents\skills\gpt-image-2\references\product-visuals\premium-studio-product.md`
+
+Template adaptation:
+
+- `concept-scene.md` is used only for foreground / subject / background layering.
+- `minimalist-mood-scene.md` is used for restraint, quiet mood, and avoiding overfilled spectacle.
+- `premium-studio-product.md` is used as a negative reference: the boot must not become a product or advertisement subject.
+- Final output should be a realistic cinematic documentary still, not concept art, not product photography, and not fantasy illustration.
+
+## Source Files Actually Used
+
+Primary shot source:
+
+- `slides/s46/shot-brief-v1.md`
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+- `visual-system/style/style-anchor-c-repair-review-v1.md`
+
+Intake:
+
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+
+## Reference Assets
+
+Primary C reference:
+
+```text
+visual-system/style/style-anchor-c-repair-v1.png
+```
+
+Role:
+
+- primary C repair reference;
+- mood / composition / texture / acceptance reference;
+- not an edit target.
+
+Style board reference:
+
+```text
+visual-system/style/style-anchor-board-v1.png
+```
+
+Role:
+
+- common style anchor reference;
+- use for palette, road-film grammar, and drift controls.
+
+Character / boot references:
+
+```text
+visual-system/characters/character-anchor-v2-codex-posture-test.png
+visual-system/characters/character-slice-walk-side-v2-from-anchor.png
+```
+
+Role:
+
+- peripheral boot / coat / lower-body identity reference;
+- not portrait target;
+- not edit target.
+
+## S-46 Target
+
+This image must answer:
+
+> 泥泞里，有星空。
+
+The viewer should first see:
+
+- real mud;
+- dark water;
+- grit, stones, shoe marks, uneven puddle edge;
+- then discover broken star reflection inside the puddle.
+
+The image should not show the character fully. The traveler is present only as an edge trace.
+
+## Rendered Prompt
+
+Use case: photorealistic-natural.
+
+Asset type: 16:9 MV trial shot still for `在路上`, `S-46`, final-four-lines muddy starlight soul shot.
+
+Primary request:
+
+Generate one 16:9 horizontal realistic cinematic documentary still for the `S-46` muddy starlight shot of the `在路上` MV. This is the soul image: a real muddy puddle at night or deep pre-dawn, with broken star points reflected in dark water. The image must say that the mud remains, but meaning appears inside it.
+
+Scene:
+
+A close-to-ground view of an irregular muddy puddle on a rough wet road or muddy ground at night or deep pre-dawn. The scene is dark, cold, and physically real. The puddle is surrounded by heavy wet mud, grit, stones, shoe marks, broken puddle edges, dirty water, small ripples, and uneven muddy texture. The water surface reflects a broken starry sky: small cold-white star points and faint scattered clusters interrupted by mud, ripples, stones, and footmarks.
+
+Main subject:
+
+The main subject is the muddy puddle and the broken star reflection in the dark water. The viewer should first read wet mud and real ground, then discover the stars inside the puddle. The star reflection must be visible enough to read, but it should be broken, partial, and physically interrupted by mud and water texture.
+
+Peripheral traveler trace:
+
+Only a small edge trace of the traveler is allowed: an old worn muddy boot at the far edge, a partial lower-leg shadow, or a small coat-hem edge. This trace should occupy under about 15% of the frame. It must not show the full person, face, portrait, or hero pose. The boot is old, road-worn, dirty, and practical, not new, not polished, not a fashion boot, and not a product closeup.
+
+Composition:
+
+Low ground-level composition. The camera is close to the mud. The puddle occupies the center or lower-center of the frame, with rough mud surrounding it. The boot or lower-body trace stays on the extreme edge only, cropped and secondary. Do not make the real sky the main image. The stars should be read as a reflection in puddle water, not as a large clean sky scene.
+
+Style:
+
+Real documentary texture + restrained poetic realism + low-place-sees-high-place composition. Low saturation. The image should belong with the accepted C repair anchor: dark blue-gray, wet black-gray, mud brown, coal black, and small cold-white star points. It should feel physically muddy and quiet, not decorative and not magical.
+
+Lighting:
+
+Night or deep pre-dawn darkness. Weak cold reflected light on water. Small cold-white star glints in the puddle. Dark and midtone areas should retain mud and water texture. No visible sun. No sunrise. No sunset. No warm horizon. No broad orange or gold glow. No firelight. No theatrical fantasy light.
+
+Mood:
+
+Quiet, low, difficult, and meaningful. The emotion is not victory and not escape. The mud is still there. The star reflection is discovered inside the mud.
+
+Strict constraints:
+
+- 16:9 horizontal frame;
+- close-to-ground composition;
+- muddy puddle as main subject;
+- broken starry sky reflected in dark water;
+- realistic wet mud texture;
+- real grit, stones, shoe marks, water ripples, and dirty puddle edges;
+- old worn boot or lower-body trace only at the frame edge;
+- boot / traveler trace under about 15% of the frame;
+- night or deep pre-dawn darkness;
+- poetic documentary realism;
+- no portrait emphasis;
+- no full person;
+- no text, no labels, no logo, no UI, no watermark.
+
+Strictly avoid:
+
+- sunrise;
+- sunset;
+- broad warm horizon;
+- beautiful postcard reflection;
+- clean mirror lake;
+- fantasy portal;
+- magic glow;
+- cosmic spectacle;
+- Milky Way as a real sky above the scene;
+- shoe advertisement;
+- product photography;
+- clean new boot;
+- fashion outdoor gear;
+- full-body traveler;
+- face or portrait;
+- heroic pose;
+- only mud with no stars;
+- only stars with no mud;
+- overly saturated colors;
+- text, symbols, watermarks, UI.
+
+Final image behavior:
+
+The viewer should first see real wet mud and dirty water. Then the broken star reflection should become clear inside the puddle. The boot or traveler trace should only prove this belongs to the traveler on the road, while the true subject remains mud, water, and broken starlight.
+
+## Reference Handling At Generation Time
+
+Current execution:
+
+- Use the rendered prompt as a standalone prompt.
+- The local reference assets were read / viewed for prompt construction.
+- The host image tool path does not provide explicit file-reference arguments in this run.
+
+If a later host path supports image references:
+
+- attach `visual-system/style/style-anchor-c-repair-v1.png` as primary C reference;
+- attach `visual-system/style/style-anchor-board-v1.png` as global style reference;
+- attach `visual-system/characters/character-slice-walk-side-v2-from-anchor.png` only as boot / lower-body reference;
+- explicitly instruct the model not to copy a full character and not to make the boot the subject.
+
+## Generation Result
+
+Generated after user authorization to continue directly:
+
+```text
+slides/s46/shot-image-v1.png
+```
+
+Review file:
+
+```text
+slides/s46/shot-review-v1.md
+```
+
+Generation mode:
+
+```text
+gpt-image-2 Skill Mode B; host image-generation tool; no Garden script.
+```
+
+## Acceptance Criteria
+
+S-46 passes only if:
+
+1. The viewer first sees real mud and water.
+2. Broken star reflection is clear enough to read.
+3. The scene is night or deep pre-dawn, not sunrise or sunset.
+4. The boot / traveler trace stays peripheral and under about `15%`.
+5. The boot does not become a product shot.
+6. The image is restrained and documentary-real, not fantasy.
+7. It truly conveys "泥泞里，有星空".
+
+## Failure Conditions
+
+Reject the generation if any of these happen:
+
+- no clear star reflection appears;
+- mud looks fake, clean, or decorative;
+- the image becomes sunrise or sunset;
+- broad warm light appears;
+- the reflection becomes a clean pretty landscape;
+- the image becomes fantasy, portal, or magic;
+- boot becomes the main subject or shoe advertisement;
+- a full person, face, or portrait appears;
+- the shot no longer feels connected to the accepted C repair.
+
+## Review File Created After Generation
+
+After generation, created:
+
+```text
+slides/s46/shot-review-v1.md
+```
+
+That review compares the generated image against:
+
+- `slides/s46/shot-brief-v1.md`;
+- this prompt/spec;
+- accepted C repair reference;
+- accepted common style anchor board;
+- accepted v2 character identity lock for boot/lower-body cues;
+- the S-46 acceptance criteria above.
diff --git a/projects/2026-06-23-在路上/slides/s46/shot-review-v1.md b/projects/2026-06-23-在路上/slides/s46/shot-review-v1.md
new file mode 100644
index 0000000..365fa6e
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s46/shot-review-v1.md
@@ -0,0 +1,158 @@
+# S-46 Shot Review V1
+
+## Status
+
+- Document status: user accepted generated candidate.
+- Execution status: S-46 image generated once.
+- Project: `在路上`.
+- Unit: `s46`.
+- Source shot: `S-46`.
+- Segment / time: Final four lines / `3:58-4:16`.
+- Style anchor: C｜泥泞里的星空.
+- Candidate image: `shot-image-v1.png`.
+- Review result: accepted by user.
+
+## Generation Trace
+
+Skill mode check command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+Skill mode check 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 · 未启用 Garden。若宿主 Agent 自带图像工具（image_generation / dalle / mcp__*image* 等）→ MODE B：把 prompt 交给宿主出图。若宿主无图像工具 → MODE C：仅产出高质量 prompt 给用户。"
+}
+```
+
+Interpretation:
+
+- Garden local generation was not enabled.
+- The host image-generation path was available.
+- This generation used `gpt-image-2` Skill Mode B: render prompt through the Skill workflow, then pass the prompt to the host image tool.
+- No Garden generation script was called.
+
+Prompt/spec used:
+
+```text
+slides/s46/shot-prompt-v1.md
+```
+
+Host default output:
+
+```text
+C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_02d1d792729fcd87016a3a54d1336881919c8457e3f17de975.png
+```
+
+Project copy:
+
+```text
+slides/s46/shot-image-v1.png
+```
+
+The original host output was left in place.
+
+## Image File Check
+
+Project image path:
+
+```text
+C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\slides\s46\shot-image-v1.png
+```
+
+File size:
+
+```text
+2488657 bytes
+```
+
+Dimensions:
+
+```text
+1672x941
+```
+
+Pixel format:
+
+```text
+Format24bppRgb
+```
+
+Aspect note:
+
+- The image is horizontal and close to 16:9.
+- It is not mathematically exact 16:9 because `1672 / 941 = 1.7768`, while exact 16:9 is `1.7777`.
+- No crop or resize was applied in this review step.
+
+## Source Alignment
+
+Primary sources checked:
+
+- `slides/s46/shot-brief-v1.md`
+- `slides/s46/shot-prompt-v1.md`
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+- `visual-system/style/style-anchor-c-repair-review-v1.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Reference asset handling:
+
+- Character and style references were read / viewed locally for prompt construction.
+- The host image tool was used as a standalone prompt path; explicit image-reference attachment was not available in this run.
+
+## Local Inspection
+
+Local visual inspection:
+
+- The first read is wet mud, dark water, puddle edge, grit, and rough ground texture.
+- Broken star reflection is clear and readable inside the puddle.
+- The scene reads as night or deep pre-dawn, not sunrise or sunset.
+- No broad warm horizon, sun, firelight, UI, text, logo, or fantasy portal is visible.
+- The boot / lower-leg trace is on the left edge and does not show a full person or face.
+- The boot is muddy and old; it does not read as a clean product pack shot.
+- The palette stays close to the accepted C repair: dark blue-gray, wet black-gray, mud brown, coal black, and cold-white star points.
+
+Watch points:
+
+- The reflected stars are very clear and dense. This helps readability, but user review should judge whether it feels too cosmic or too clean for the intended harsh documentary tone.
+- The boot / lower-leg trace on the left edge has noticeable presence. It appears secondary to the puddle, but it is close to the upper limit of the intended peripheral role.
+
+External vision analysis:
+
+- Not used for this review.
+- This review relies on local visual inspection and local file metadata only.
+
+## Acceptance Criteria Review
+
+| Criterion | Assessment | Notes |
+| --- | --- | --- |
+| Viewer first sees real mud and water | pass for review | Mud, puddle edge, and wet ground dominate the frame. |
+| Broken star reflection is clear enough | pass for review | Star points are readable inside the dark water. |
+| Night or deep pre-dawn, not sunrise/sunset | pass for review | No visible sun or warm horizon. |
+| Boot / traveler trace peripheral under about 15% | mostly pass for review | Boot and leg stay at left edge, but their presence is noticeable. |
+| Boot does not become a product shot | pass for review | Boot is dirty and secondary, not clean or advertised. |
+| Restrained and documentary-real, not fantasy | mostly pass for review | Mud is real; star density is the main watch point. |
+| Conveys "泥泞里，有星空" | pass for review | The core image is readable. |
+
+## Decision
+
+User accepted:
+
+```text
+slides/s46/shot-image-v1.png
+```
+
+Next gated shot can begin from `S-47`.
diff --git a/projects/2026-06-23-在路上/slides/s47/shot-brief-v1.md b/projects/2026-06-23-在路上/slides/s47/shot-brief-v1.md
new file mode 100644
index 0000000..2dd242a
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s47/shot-brief-v1.md
@@ -0,0 +1,194 @@
+# S-47 Shot Brief V1
+
+## Status
+
+- Document status: created for direct S-47 generation.
+- Execution status: prompt/spec created and used directly; S-47 image accepted by user.
+- Project: `在路上`.
+- Unit: `s47`.
+- Source shot: `S-47`.
+- Segment / time: Outro / `4:16-4:54`.
+- Trial role: fourth shot in the five-shot visual test package.
+- Style anchor: D｜夜路小火.
+
+## Purpose
+
+S-47 tests whether the accepted C soul image can transition into D's restrained warmth.
+
+Core test sentence:
+
+> 星光变成火光，但人还没有到达终点。
+
+This shot is not a camping scene, a reunion, or a victory pause. It is a brief roadside confirmation that the traveler can continue.
+
+## Source Files Actually Used
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+
+Accepted image references:
+
+- `visual-system/style/style-anchor-board-v1.png`
+- `intake/imgs/D.夜路篝火型蓝图.png`
+- `visual-system/characters/character-anchor-v2-codex-posture-test.png`
+- `visual-system/characters/character-slice-walk-back-v2-from-anchor.png`
+- `slides/s46/shot-image-v1.png`
+- `slides/s46/shot-review-v1.md`
+
+Primary intake:
+
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+
+Skill references:
+
+- `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`
+- `C:\Users\wangq\.codex\skills\.system\imagegen\references\prompting.md`
+
+## Source Meaning
+
+From the shot skeleton:
+
+- S-47 is the start of the outro.
+- It lifts from the muddy starlight image into a real night road.
+- The traveler pauses near a small fire.
+- It converts star reflection into firelight.
+- It begins the final "far away there are others" confirmation without showing a reunion.
+
+From the five-shot trial plan:
+
+- It must show a night road after the muddy starlight moment.
+- It must include a small local roadside fire.
+- The traveler must be back view or side-back view, not portrait.
+- Wet dark road / dark wilderness texture must remain visible.
+- Warm fire must stay a small local signal.
+
+Acceptance sentence:
+
+> The image should say "he can continue", not "he has arrived".
+
+## Composition Target
+
+Canvas:
+
+- 16:9 horizontal cinematic still.
+
+Shot scale:
+
+- Medium-wide road-film frame.
+- Environment and road remain larger than the person.
+- Traveler about 1/8 to 1/6 of the frame, not a portrait.
+- Fire about 3% to 6% of the frame, visible but small.
+
+Viewpoint:
+
+- Road-level or slightly low roadside perspective.
+- The road should recede into dark distance.
+- Foreground can include wet mud, shallow puddles, and cold residual starlight.
+- Firelight should be reflected only locally on mud and water.
+
+Subject placement:
+
+- Traveler in back or side-back view near a small roadside fire.
+- Body should remain upright, chest open, and able to continue.
+- He may pause, but he should not settle into camp.
+
+## Character Requirements
+
+Use the accepted character identity lock:
+
+- 50-year-old male traveler, about 180cm tall.
+- Dark slightly longer messy hair; face can remain unseen.
+- Dark gray / dark khaki / coal-black long road-worn coat.
+- Old practical backpack.
+- Old dark pants and old walking boots.
+- Tired but dignified; not collapsed.
+
+Critical posture lock:
+
+- no bent waist;
+- no hunched back;
+- no rounded collapsed shoulders;
+- no bowed defeated head;
+- no frail, broken body.
+
+## D Anchor Requirements
+
+Must preserve:
+
+- small fire;
+- dark road;
+- distant dark wilderness;
+- wet ground;
+- restrained warmth;
+- documentary road-film texture.
+
+Must avoid:
+
+- camping advertisement;
+- tent, chair, picnic setup, cookware, lifestyle props;
+- large campfire;
+- group gathering;
+- clear reunion;
+- triumphant rest;
+- cozy lifestyle scene;
+- victory poster.
+
+## Palette And Light
+
+Base:
+
+- deep blue-black night;
+- wet dark blue-gray;
+- coal black;
+- mud brown;
+- dark khaki.
+
+Warmth:
+
+- small warm orange fire only;
+- local glow on the traveler's coat edge, boots, mud, and puddles;
+- no broad golden horizon.
+
+Avoid:
+
+- sunrise;
+- sunset;
+- bright daylight;
+- neon/cyberpunk;
+- fantasy magic glow;
+- over-warm cinematic camping color.
+
+## Acceptance Criteria
+
+S-47 passes only if:
+
+1. The first read is night road and wet ground, not a camping portrait.
+2. The fire is small and local.
+3. The traveler reads as the same road-worn figure through coat, backpack, boots, and back/side-back silhouette.
+4. The posture is tired but upright and capable.
+5. The image feels like continuation, not arrival.
+6. There is no group, tent, cozy camp setup, or victory tone.
+7. It can sit after S-46 and before S-51 without changing the MV's visual language.
+
+## Next Files
+
+```text
+slides/s47/shot-prompt-v1.md
+slides/s47/shot-image-v1.png
+slides/s47/shot-review-v1.md
+```
diff --git a/projects/2026-06-23-在路上/slides/s47/shot-image-v1.png b/projects/2026-06-23-在路上/slides/s47/shot-image-v1.png
new file mode 100644
index 0000000..158f38c
Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s47/shot-image-v1.png differ
diff --git a/projects/2026-06-23-在路上/slides/s47/shot-prompt-v1.md b/projects/2026-06-23-在路上/slides/s47/shot-prompt-v1.md
new file mode 100644
index 0000000..cf5f2be
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s47/shot-prompt-v1.md
@@ -0,0 +1,194 @@
+# S-47 Shot Prompt V1
+
+## Status
+
+- Document status: created for direct generation.
+- Execution status: used once for Mode B host image generation; candidate image accepted by user.
+- Project: `在路上`.
+- Unit: `s47`.
+- Source shot: `S-47`.
+- Segment / time: Outro / `4:16-4:54`.
+- Style anchor: D｜夜路小火.
+- Purpose: generate the fourth trial image for the C-to-D night-road-fire transition.
+- Generated candidate: `shot-image-v1.png`.
+
+## User Gate
+
+User accepted S-46 and requested the next shot.
+
+The current trial workflow skips intermediate prompt review. The user reviews final generated images.
+
+## Skill Mode Check
+
+Command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+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 · 未启用 Garden。若宿主 Agent 自带图像工具（image_generation / dalle / mcp__*image* 等）→ MODE B：把 prompt 交给宿主出图。若宿主无图像工具 → MODE C：仅产出高质量 prompt 给用户。"
+}
+```
+
+Interpretation:
+
+- Garden local generation is not enabled.
+- The host image-generation path is available in this environment.
+- Use `gpt-image-2` Skill Mode B: render the prompt through the Skill workflow, then pass it to the host image tool.
+- Do not call `scripts/generate.js`.
+
+## Skill Template / Reference Basis
+
+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`
+- `C:\Users\wangq\.codex\skills\.system\imagegen\SKILL.md`
+- `C:\Users\wangq\.codex\skills\.system\imagegen\references\prompting.md`
+
+Template selection:
+
+- Primary category: `scenes-and-illustrations`.
+- Main template logic: `concept-scene.md` for cinematic road-film composition with foreground / midground / background layers.
+- Secondary restraint: `minimalist-mood-scene.md` for reduced elements and emotional restraint.
+- Host image tool prompt structure: imagegen shared prompt schema.
+
+## Project Sources Actually Used
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+- `slides/five-shot-trial-plan-v1.md`
+- `slides/s46/shot-review-v1.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+
+Intake sources:
+
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+
+Viewed local references:
+
+- `visual-system/style/style-anchor-board-v1.png`
+- `intake/imgs/D.夜路篝火型蓝图.png`
+- `visual-system/characters/character-anchor-v2-codex-posture-test.png`
+- `visual-system/characters/character-slice-walk-back-v2-from-anchor.png`
+- `slides/s46/shot-image-v1.png`
+
+Reference roles:
+
+- D blueprint / style board: style and composition reference.
+- Character anchor / back-walking slice: identity and silhouette reference.
+- S-46 image: previous accepted shot, used as sequence continuity reference.
+
+The host image tool is used as a standalone prompt path. Explicit image-reference attachment is not available in this run.
+
+## S-47 Target
+
+Shot function:
+
+- transition from C muddy starlight to D night-road fire;
+- make star reflection become firelight in a realistic way;
+- prove warmth without camping, group reunion, or victory.
+
+Core acceptance sentence:
+
+> The image should say "he can continue", not "he has arrived".
+
+## Final Rendered Prompt For Host Image Tool
+
+Use case: photorealistic-natural
+
+Asset type: 16:9 MV trial shot still for `在路上`, `S-47`, outro C-to-D transition.
+
+Primary request:
+
+Generate one 16:9 horizontal realistic cinematic documentary still for the `S-47` night-road-fire transition in the `在路上` MV. This shot happens immediately after the accepted muddy starlight puddle moment. The visual idea is that the cold star reflection has become a small real roadside fire, but the road continues.
+
+Scene/backdrop:
+
+A dark wet wilderness road at night or deep pre-dawn. The road is muddy, old, uneven, and wet, with shallow puddles and rough ground texture. The road recedes into deep blue-black distance. The world is larger than the person. Keep the landscape quiet and sparse, not scenic tourism.
+
+Subject:
+
+A 50-year-old male traveler, about 180cm tall, seen only from back or side-back view. He wears a dark gray / dark khaki / coal-black long road-worn coat, old dark pants, old practical backpack, and old walking boots. His longer messy dark hair silhouette may be visible, but the face is not the subject. He pauses beside a very small roadside fire, body turned toward the road as if ready to continue. His posture is tired but upright and dignified, chest open, spine straight, shoulders not collapsed.
+
+Composition/framing:
+
+Medium-wide road-film frame. Environment first, road second, traveler third. The traveler should occupy about 1/8 to 1/6 of the frame, not a portrait. The small fire should occupy only about 3% to 6% of the frame. Place the fire near the roadside, not in the center as a camp scene. Let the wet road and dark distance dominate the frame. Include a few small puddles catching cold night reflections and a restrained warm fire reflection in nearby mud and water.
+
+Lighting/mood:
+
+Deep blue-black night with readable dark depth. The only warm light is the small local orange fire. It creates a modest glow on the traveler's coat edge, boots, nearby mud, and puddles. No broad golden horizon, no sunrise, no sunset, no daylight. Mood is quiet, restrained, tired, and continuing.
+
+Style/medium:
+
+Photorealistic cinematic documentary still, realistic road texture, wet mud, worn fabric, practical outdoor wear, natural firelight, restrained poetic realism. It should feel like a road-film frame, not fantasy art, not a lifestyle camping advertisement, not a poster.
+
+Character invariants:
+
+Keep the traveler road-worn and capable. He may be tired, but do not make him physically broken. No bent waist, no hunched back, no rounded collapsed shoulders, no bowed defeated head, no frail body. Same visual identity cues: long dark coat, old backpack, old boots, restrained middle-aged traveler silhouette.
+
+Required emotional meaning:
+
+The image should say: he can continue. It must not say: he has arrived, he is celebrating, or he has joined a group.
+
+Avoid:
+
+No tent, no camping chairs, no cookware, no picnic setup, no cozy campsite, no lifestyle advertising, no large campfire, no group gathering, no visible reunion, no cheering, no victory pose, no hero stance, no front-facing portrait, no arms-open inspiration posture, no fashion travel photography, no homeless stereotype, no drunk stereotype, no sci-fi, no cyberpunk, no neon, no UI, no code, no text, no logos, no watermark, no fantasy magic glow, no sunrise, no sunset, no broad warm horizon.
+
+Output intent:
+
+One still image candidate for `slides/s47/shot-image-v1.png`.
+
+## Generation Result
+
+Host default output:
+
+```text
+C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0301a08a35fd2c33016a3a580483248190aaa61db4afc34c0e.png
+```
+
+Project copy:
+
+```text
+slides/s47/shot-image-v1.png
+```
+
+## Acceptance Criteria To Use In Review
+
+S-47 passes only if:
+
+1. The first read is night road and wet ground, not a camping portrait.
+2. The fire is small and local.
+3. The traveler reads as the same road-worn figure through coat, backpack, boots, and back/side-back silhouette.
+4. The posture is tired but upright and capable.
+5. The image feels like continuation, not arrival.
+6. There is no group, tent, cozy camp setup, or victory tone.
+7. It can sit after S-46 and before S-51 without changing the MV's visual language.
+
+## Review File Created After Generation
+
+```text
+slides/s47/shot-review-v1.md
+```
diff --git a/projects/2026-06-23-在路上/slides/s47/shot-review-v1.md b/projects/2026-06-23-在路上/slides/s47/shot-review-v1.md
new file mode 100644
index 0000000..e588a0d
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s47/shot-review-v1.md
@@ -0,0 +1,158 @@
+# S-47 Shot Review V1
+
+## Status
+
+- Document status: user accepted generated candidate.
+- Execution status: S-47 image generated once.
+- Project: `在路上`.
+- Unit: `s47`.
+- Source shot: `S-47`.
+- Segment / time: Outro / `4:16-4:54`.
+- Style anchor: D｜夜路小火.
+- Candidate image: `shot-image-v1.png`.
+- Review result: accepted by user.
+
+## Generation Trace
+
+Skill mode check command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+Skill mode check 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 · 未启用 Garden。若宿主 Agent 自带图像工具（image_generation / dalle / mcp__*image* 等）→ MODE B：把 prompt 交给宿主出图。若宿主无图像工具 → MODE C：仅产出高质量 prompt 给用户。"
+}
+```
+
+Interpretation:
+
+- Garden local generation was not enabled.
+- The host image-generation path was available.
+- This generation used `gpt-image-2` Skill Mode B: render prompt through the Skill workflow, then pass the prompt to the host image tool.
+- No Garden generation script was called.
+
+Prompt/spec used:
+
+```text
+slides/s47/shot-prompt-v1.md
+```
+
+Host default output:
+
+```text
+C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0301a08a35fd2c33016a3a580483248190aaa61db4afc34c0e.png
+```
+
+Project copy:
+
+```text
+slides/s47/shot-image-v1.png
+```
+
+The original host output was left in place.
+
+## Image File Check
+
+Project image path:
+
+```text
+C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\slides\s47\shot-image-v1.png
+```
+
+File size:
+
+```text
+1834279 bytes
+```
+
+Dimensions:
+
+```text
+1672x941
+```
+
+Pixel format:
+
+```text
+Format24bppRgb
+```
+
+Aspect note:
+
+- The image is horizontal and close to 16:9.
+- It is not mathematically exact 16:9 because `1672 / 941 = 1.7768`, while exact 16:9 is `1.7777`.
+- No crop or resize was applied in this review step.
+
+## Source Alignment
+
+Primary sources checked:
+
+- `slides/s47/shot-brief-v1.md`
+- `slides/s47/shot-prompt-v1.md`
+- `slides/s46/shot-review-v1.md`
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Reference asset handling:
+
+- Character and style references were read / viewed locally for prompt construction.
+- The host image tool was used as a standalone prompt path; explicit image-reference attachment was not available in this run.
+
+## Local Inspection
+
+Local visual inspection:
+
+- The image reads as a dark wilderness night road with wet mud, puddles, and cold blue-black depth.
+- A small local fire sits beside the road and provides restrained warm orange light.
+- The traveler is back view, wearing a long dark coat and backpack, and stands near the fire while facing the continuing road.
+- The posture is upright and capable; it does not show a hunched, collapsed, or defeated body.
+- No tent, chairs, cookware, picnic setup, group gathering, reunion, UI, text, logo, sci-fi, or neon is visible.
+- The road continues into darkness, so the image still reads as "continue" rather than "arrived".
+- Firelight is local and touches nearby mud, puddles, boots, and coat edge without turning the whole scene warm.
+
+Watch points:
+
+- The traveler is near the upper size limit for S-47, roughly around the intended 1/8 to 1/6 band. User review should judge whether this is still environment-led enough.
+- The starry sky and distant fire points are readable. This helps the C-to-D transition, but user review should judge whether it previews S-48/S-51 too strongly.
+
+External vision analysis:
+
+- Not used for this review.
+- This review relies on local visual inspection and local file metadata only.
+
+## Acceptance Criteria Review
+
+| Criterion | Assessment | Notes |
+| --- | --- | --- |
+| First read is night road and wet ground, not camping portrait | pass for review | Wet road and dark distance dominate with no camp setup. |
+| Fire is small and local | pass for review | Fire stays low and modest; warmth is localized. |
+| Traveler reads through coat, backpack, boots, back silhouette | pass for review | Identity cues are clear from the back view. |
+| Posture tired but upright and capable | pass for review | The body is standing and road-facing, not collapsed. |
+| Feels like continuation, not arrival | pass for review | The road continues into darkness. |
+| No group, tent, cozy camp setup, or victory tone | pass for review | None visible. |
+| Fits after S-46 and before S-51 | mostly pass for review | Starry sky and distant fire points create continuity; watch the strength of the far-fire hint. |
+
+## Decision
+
+User accepted:
+
+```text
+slides/s47/shot-image-v1.png
+```
+
+Next gated shot can begin from `S-51`.
diff --git a/projects/2026-06-23-在路上/slides/s51/shot-brief-v1.md b/projects/2026-06-23-在路上/slides/s51/shot-brief-v1.md
new file mode 100644
index 0000000..9f7659b
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s51/shot-brief-v1.md
@@ -0,0 +1,188 @@
+# S-51 Shot Brief V1
+
+## Status
+
+- Document status: created for direct S-51 generation.
+- Execution status: prompt/spec created and used directly; S-51 image accepted by user.
+- Project: `在路上`.
+- Unit: `s51`.
+- Source shot: `S-51`.
+- Segment / time: Outro / `4:16-4:54`.
+- Trial role: fifth and final shot in the five-shot visual test package.
+- Style anchor: D｜夜路小火 / 远方火点.
+
+## Purpose
+
+S-51 tests the ending restraint of the MV.
+
+Core test sentence:
+
+> 远方也有人在走，所以他继续在路上。
+
+This shot is not a resolution, victory, reunion, or arrival. It is an unfinished final wide shot: the road remains larger than the traveler, but the far fires prove he is not completely alone.
+
+## Source Files Actually Used
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+
+Accepted image references:
+
+- `visual-system/style/style-anchor-board-v1.png`
+- `intake/imgs/D.夜路篝火型蓝图.png`
+- `visual-system/characters/character-slice-walk-back-v2-from-anchor.png`
+- `slides/s47/shot-image-v1.png`
+- `slides/s47/shot-review-v1.md`
+
+Primary intake:
+
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+
+Skill references:
+
+- `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`
+- `C:\Users\wangq\.codex\skills\.system\imagegen\references\prompting.md`
+
+## Source Meaning
+
+From the shot skeleton:
+
+- S-51 is the outro tail and final resonance image.
+- It is a very wide shot.
+- Night road, several distant fires, sky, wilderness, and world must all exist at once.
+- The traveler becomes almost a small point inside the road.
+- The ending does not solve the journey; it leaves "still on the road".
+
+From the expanded key-shot source:
+
+- It is not a victory summary.
+- It is not a meeting, hug, or team gathering.
+- It confirms that far away there are others, then keeps the traveler on the road.
+- The viewer should feel unfinished but less alone.
+
+Acceptance sentence:
+
+> It should feel unfinished but less alone: the traveler is still on the road, and far away there are other fires.
+
+## Composition Target
+
+Canvas:
+
+- 16:9 horizontal cinematic still.
+
+Shot scale:
+
+- Very wide night-road / wilderness-road frame.
+- Traveler about 1/20 to 1/12 of the frame.
+- Road, sky, and wilderness must occupy most of the image.
+- Person should be readable mainly as a small walking silhouette, not a portrait.
+
+Viewpoint:
+
+- Wide road-film landscape view.
+- Road starts from near foreground or lower frame and runs into dark distance.
+- The traveler walks away along or beside the road.
+- One small near/behind fire can be present, but it must not dominate.
+- One or two distant weak fire points should be visible far ahead.
+
+## Character Requirements
+
+Use only wide-shot identity cues:
+
+- small back-view walking figure;
+- long dark road-worn coat;
+- old backpack;
+- old walking boots only if leg detail is visible;
+- upright, capable walking posture.
+
+Do not show:
+
+- face;
+- emotional close-up;
+- heroic backshot centered like a poster;
+- arms-open posture;
+- bent waist or collapsed gait.
+
+## D Anchor Requirements
+
+Must preserve:
+
+- deep blue-black night;
+- dark muddy or wilderness road;
+- one small local fire or residual near fire;
+- one or two distant weak fire points;
+- restrained warmth;
+- real documentary texture;
+- road/world larger than person.
+
+Must avoid:
+
+- camping site;
+- tent, chairs, cookware, picnic setup;
+- group reunion;
+- visible crowd;
+- victory or arrival image;
+- large theatrical firelight;
+- inspirational poster backshot;
+- final destination feeling.
+
+## Palette And Light
+
+Base:
+
+- deep blue-black night;
+- wet dark blue-gray;
+- coal black;
+- mud brown;
+- dark khaki;
+- cold gray.
+
+Warmth:
+
+- very small warm-orange fire points only;
+- near fire light should be local and low;
+- distant fires should be weak, not celebratory.
+
+Avoid:
+
+- sunrise or sunset;
+- broad warm horizon;
+- daylight;
+- neon/cyberpunk;
+- fantasy glow;
+- over-bright stars.
+
+## Acceptance Criteria
+
+S-51 passes only if:
+
+1. The first read is road, night, sky, and wilderness, not a person or campfire.
+2. The traveler is very small, about 1/20 to 1/12 of the frame.
+3. The road is still more important than the person.
+4. Far fires read as "there are others on the road" without showing a group.
+5. Warmth is real but restrained.
+6. No reunion, no victory, no camping, no final destination feeling.
+7. The ending feels unfinished but less alone.
+
+## Next Files
+
+```text
+slides/s51/shot-prompt-v1.md
+slides/s51/shot-image-v1.png
+slides/s51/shot-review-v1.md
+```
diff --git a/projects/2026-06-23-在路上/slides/s51/shot-image-v1.png b/projects/2026-06-23-在路上/slides/s51/shot-image-v1.png
new file mode 100644
index 0000000..f4faf5a
Binary files /dev/null and b/projects/2026-06-23-在路上/slides/s51/shot-image-v1.png differ
diff --git a/projects/2026-06-23-在路上/slides/s51/shot-prompt-v1.md b/projects/2026-06-23-在路上/slides/s51/shot-prompt-v1.md
new file mode 100644
index 0000000..7452772
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s51/shot-prompt-v1.md
@@ -0,0 +1,194 @@
+# S-51 Shot Prompt V1
+
+## Status
+
+- Document status: created for direct generation.
+- Execution status: used once for Mode B host image generation; candidate image accepted by user.
+- Project: `在路上`.
+- Unit: `s51`.
+- Source shot: `S-51`.
+- Segment / time: Outro / `4:16-4:54`.
+- Style anchor: D｜夜路小火 / 远方火点.
+- Purpose: generate the fifth and final trial image for the ending wide shot.
+- Generated candidate: `shot-image-v1.png`.
+
+## User Gate
+
+User accepted S-47 and requested the next shot.
+
+The current trial workflow skips intermediate prompt review. The user reviews final generated images.
+
+## Skill Mode Check
+
+Command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+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 · 未启用 Garden。若宿主 Agent 自带图像工具（image_generation / dalle / mcp__*image* 等）→ MODE B：把 prompt 交给宿主出图。若宿主无图像工具 → MODE C：仅产出高质量 prompt 给用户。"
+}
+```
+
+Interpretation:
+
+- Garden local generation is not enabled.
+- The host image-generation path is available in this environment.
+- Use `gpt-image-2` Skill Mode B: render the prompt through the Skill workflow, then pass it to the host image tool.
+- Do not call `scripts/generate.js`.
+
+## Skill Template / Reference Basis
+
+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`
+- `C:\Users\wangq\.codex\skills\.system\imagegen\SKILL.md`
+- `C:\Users\wangq\.codex\skills\.system\imagegen\references\prompting.md`
+
+Template selection:
+
+- Primary category: `scenes-and-illustrations`.
+- Main template logic: `concept-scene.md` for wide road-film landscape layers.
+- Secondary restraint: `minimalist-mood-scene.md` for reduced elements, large negative space, and emotional restraint.
+- Host image tool prompt structure: imagegen shared prompt schema.
+
+## Project Sources Actually Used
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+- `slides/slides.md`
+- `slides/five-shot-trial-plan-v1.md`
+- `slides/s47/shot-review-v1.md`
+
+Visual system:
+
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+
+Intake sources:
+
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+
+Viewed local references:
+
+- `visual-system/style/style-anchor-board-v1.png`
+- `intake/imgs/D.夜路篝火型蓝图.png`
+- `visual-system/characters/character-slice-walk-back-v2-from-anchor.png`
+- `slides/s47/shot-image-v1.png`
+
+Reference roles:
+
+- D blueprint / style board: D warmth and restrained night-road atmosphere.
+- Character walk-back slice: small back-view identity cues.
+- S-47 image: immediately previous accepted transition shot.
+
+The host image tool is used as a standalone prompt path. Explicit image-reference attachment is not available in this run.
+
+## S-51 Target
+
+Shot function:
+
+- final wide shot of the MV;
+- road/world returns to the center;
+- far fires confirm that others exist, without showing a gathering;
+- the ending remains unfinished but less alone.
+
+Core acceptance sentence:
+
+> It should feel unfinished but less alone: the traveler is still on the road, and far away there are other fires.
+
+## Final Rendered Prompt For Host Image Tool
+
+Use case: photorealistic-natural
+
+Asset type: 16:9 MV trial shot still for `在路上`, `S-51`, final outro wide shot.
+
+Primary request:
+
+Generate one 16:9 horizontal realistic cinematic documentary still for the `S-51` final wide shot of the `在路上` MV. This is the closing resonance image: a vast dark night road, a tiny traveler still walking, and a few weak fires far away. The ending should feel unfinished but less alone.
+
+Scene/backdrop:
+
+A vast dark wilderness road at night or deep pre-dawn. The ground is muddy, old, uneven, and wet, with faint puddles and rough earth texture. The road begins near the lower frame and recedes deep into the distance. The sky and wilderness are large and quiet, deep blue-black with sparse cold star points or weak cold night glow. The world must feel bigger than the person, and the road must feel longer than the person.
+
+Subject:
+
+A tiny 50-year-old male traveler walking away along the road. He is visible only as a small back-view silhouette with a long dark road-worn coat, old backpack, and restrained upright walking posture. No face, no expression, no close detail. He is still moving forward, not standing heroically, not celebrating, not arriving.
+
+Composition/framing:
+
+Very wide road-film landscape frame. Environment first, road second, distant fire points third, traveler last. The traveler should occupy only about 1/20 to 1/12 of the frame. Put the traveler low in the frame or slightly off-center, small enough to nearly merge with the road but still readable as a walking figure. Include one very small near or behind fire and one or two weak distant fire points far ahead in the darkness. Do not place fires as a camp scene; use them as sparse signals along the road. Keep the road, sky, and wilderness dominant.
+
+Lighting/mood:
+
+Deep blue-black night with readable dark depth. Warmth comes only from tiny orange fire points: a very restrained local near-fire glow and weaker distant fires. No broad warm horizon, no sunrise, no sunset, no daylight. Mood is quiet, unfinished, restrained, and less alone, not triumphant.
+
+Style/medium:
+
+Photorealistic cinematic documentary road-film still, realistic wet mud, dark wilderness texture, restrained poetic realism, natural low firelight, cold night atmosphere. It should feel like a final frame of a road documentary, not fantasy art, not a camping advertisement, not a motivational poster.
+
+Character invariants:
+
+Keep the traveler road-worn and capable. At this scale, identity survives only through silhouette: long dark coat, old backpack, walking direction, and upright gait. No hunched back, no bent waist, no collapsed shoulders, no bowed defeated head, no heroic arms-open pose.
+
+Required emotional meaning:
+
+The image should say: he is still on the road, and far away there are others on the road too. It must not say: he has arrived, he has won, he is joining a group, or the story is solved.
+
+Avoid:
+
+No tent, no camping chairs, no cookware, no picnic setup, no cozy campsite, no visible group, no reunion, no team gathering, no cheering, no victory pose, no central heroic backshot, no final destination gate, no large theatrical campfire, no warm sunset horizon, no sunrise, no daylight, no travel advertisement, no fantasy magic glow, no sci-fi, no cyberpunk, no neon, no UI, no code, no text, no logos, no watermark.
+
+Output intent:
+
+One still image candidate for `slides/s51/shot-image-v1.png`.
+
+## Generation Result
+
+Host default output:
+
+```text
+C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0f4f5d34b9230f8d016a3a5bfa40b88191a5737a2fb84a499f.png
+```
+
+Project copy:
+
+```text
+slides/s51/shot-image-v1.png
+```
+
+## Acceptance Criteria To Use In Review
+
+S-51 passes only if:
+
+1. The first read is road, night, sky, and wilderness, not a person or campfire.
+2. The traveler is very small, about 1/20 to 1/12 of the frame.
+3. The road is still more important than the person.
+4. Far fires read as "there are others on the road" without showing a group.
+5. Warmth is real but restrained.
+6. No reunion, no victory, no camping, no final destination feeling.
+7. The ending feels unfinished but less alone.
+
+## Review File Created After Generation
+
+```text
+slides/s51/shot-review-v1.md
+```
diff --git a/projects/2026-06-23-在路上/slides/s51/shot-review-v1.md b/projects/2026-06-23-在路上/slides/s51/shot-review-v1.md
new file mode 100644
index 0000000..8229653
--- /dev/null
+++ b/projects/2026-06-23-在路上/slides/s51/shot-review-v1.md
@@ -0,0 +1,158 @@
+# S-51 Shot Review V1
+
+## Status
+
+- Document status: user accepted generated candidate.
+- Execution status: S-51 image generated once.
+- Project: `在路上`.
+- Unit: `s51`.
+- Source shot: `S-51`.
+- Segment / time: Outro / `4:16-4:54`.
+- Style anchor: D｜夜路小火 / 远方火点.
+- Candidate image: `shot-image-v1.png`.
+- Review result: accepted by user.
+
+## Generation Trace
+
+Skill mode check command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+Skill mode check 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 · 未启用 Garden。若宿主 Agent 自带图像工具（image_generation / dalle / mcp__*image* 等）→ MODE B：把 prompt 交给宿主出图。若宿主无图像工具 → MODE C：仅产出高质量 prompt 给用户。"
+}
+```
+
+Interpretation:
+
+- Garden local generation was not enabled.
+- The host image-generation path was available.
+- This generation used `gpt-image-2` Skill Mode B: render prompt through the Skill workflow, then pass the prompt to the host image tool.
+- No Garden generation script was called.
+
+Prompt/spec used:
+
+```text
+slides/s51/shot-prompt-v1.md
+```
+
+Host default output:
+
+```text
+C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0f4f5d34b9230f8d016a3a5bfa40b88191a5737a2fb84a499f.png
+```
+
+Project copy:
+
+```text
+slides/s51/shot-image-v1.png
+```
+
+The original host output was left in place.
+
+## Image File Check
+
+Project image path:
+
+```text
+C:\Users\wangq\Documents\Codex\video-workbench\projects\2026-06-23-在路上\slides\s51\shot-image-v1.png
+```
+
+File size:
+
+```text
+1909982 bytes
+```
+
+Dimensions:
+
+```text
+1672x941
+```
+
+Pixel format:
+
+```text
+Format24bppRgb
+```
+
+Aspect note:
+
+- The image is horizontal and close to 16:9.
+- It is not mathematically exact 16:9 because `1672 / 941 = 1.7768`, while exact 16:9 is `1.7777`.
+- No crop or resize was applied in this review step.
+
+## Source Alignment
+
+Primary sources checked:
+
+- `slides/s51/shot-brief-v1.md`
+- `slides/s51/shot-prompt-v1.md`
+- `slides/s47/shot-review-v1.md`
+- `intake/阶段五-镜头清单骨架.md`
+- `intake/阶段五-核心单镜.md`
+- `visual-system/visual-system.md`
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/style/style-anchor-board-review-v1.md`
+- `slides/five-shot-trial-plan-v1.md`
+
+Reference asset handling:
+
+- Character and style references were read / viewed locally for prompt construction.
+- The host image tool was used as a standalone prompt path; explicit image-reference attachment was not available in this run.
+
+## Local Inspection
+
+Local visual inspection:
+
+- The first read is a vast night wilderness road, dark sky, muddy ground, and distant orange fire points.
+- The traveler is very small, back-facing, and walking away along the road.
+- The near fire is small and low; distant fires are weak signals in the dark.
+- Road, sky, and wilderness are visually more important than the traveler.
+- No tent, chairs, cookware, picnic setup, visible group, reunion, UI, text, logo, sci-fi, or neon is visible.
+- The image keeps the "still on the road" feeling; it does not show arrival or victory.
+- Warmth is present but limited to small fire points and local ground reflection.
+
+Watch points:
+
+- The star field is clear and visually attractive. User review should judge whether it stays documentary enough or becomes too pretty.
+- The traveler is small enough for the final wide; user review should judge whether he is still readable enough as the accepted road-worn figure.
+
+External vision analysis:
+
+- Not used for this review.
+- This review relies on local visual inspection and local file metadata only.
+
+## Acceptance Criteria Review
+
+| Criterion | Assessment | Notes |
+| --- | --- | --- |
+| First read is road, night, sky, and wilderness | pass for review | Environment dominates. |
+| Traveler about 1/20 to 1/12 of frame | pass for review | The figure is very small and not portrait-like. |
+| Road remains more important than person | pass for review | The road leads through the frame and carries the image. |
+| Far fires suggest others without group | pass for review | Fire points are visible; no people or gathering shown. |
+| Warmth real but restrained | pass for review | Firelight stays small and local. |
+| No reunion, victory, camping, final destination | pass for review | None visible. |
+| Ending feels unfinished but less alone | pass for review | The road continues, with weak distant fires ahead. |
+
+## Decision
+
+User accepted:
+
+```text
+slides/s51/shot-image-v1.png
+```
+
+The five-shot visual trial can be marked as image-generated and user-accepted.
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 69911d3..64352dc 100644
--- a/projects/2026-06-23-在路上/slides/slides.md
+++ b/projects/2026-06-23-在路上/slides/slides.md
@@ -27,7 +27,7 @@ Source:
 
 | Unit | Source Shot | Segment / Time | Function | Style Anchor | Batch | Status | Notes |
 | --- | --- | --- | --- | --- | --- | --- | --- |
-| s01 | S-01 | Prelude / 0:00-0:16 | World appears before person; cold frosted muddy road, no person. | A | anchor-test, priority | not started | Tests main landscape grammar. |
+| s01 | S-01 | Prelude / 0:00-0:16 | World appears before person; cold frosted muddy road, no person. | A | anchor-test, priority | accepted | `shot-image-v1.png` accepted by user. |
 | s02 | S-02 | Prelude / 0:00-0:16 | Tiny traveler enters the world along the muddy road. | A | deferred | not started | Useful after character anchor exists. |
 | s03 | S-03 | Prelude / 0:00-0:16 | Old boot and backpack detail connects body to road. | character detail | deferred | not started | Avoid shoe-ad framing. |
 | s04 | S-04 | Section A / 0:16-0:32 | Cold morning frost and long road. | A | deferred | not started | Person absent or very small. |
@@ -36,7 +36,7 @@ Source:
 | s07 | S-07 | Section A / 0:16-0:32 | Wind/sand presses against tiny traveler. | A | deferred | not started | Not heroic; person is being pressed by weather. |
 | s08 | S-08 | Section B / 0:32-1:05 | Mud on boot sole and dragging footstep. | character detail | deferred | not started | Physical weight and mud. |
 | s09 | S-09 | Section B / 0:32-1:05 | Traveler under large bridge/shadow, pressure begins. | B | priority | not started | Leads into steel canyon. |
-| s10 | S-10 | Section B / 0:32-1:05 | Steel canyon / city pressure. | B | anchor-test, expanded, priority | not started | Expanded in `阶段五-核心单镜.md`. No UI/code/sci-fi. |
+| s10 | S-10 | Section B / 0:32-1:05 | Steel canyon / city pressure. | B | anchor-test, expanded, priority | accepted | `shot-image-v1.png` accepted by user. No UI/code/sci-fi. |
 | s11 | S-11 | Section B / 0:32-1:05 | City edge, glass/cold wall, no home. | B | deferred | not started | Avoid polished city promo. |
 | s12 | S-12 | Section B / 0:32-1:05 | Rust, dripping water, time wear. | B | deferred | not started | Environment carries the passage of time. |
 | s13 | S-13 | Section B / 0:32-1:05 | Calloused hand near backpack strap or cold rail. | character detail | deferred | not started | Body detail replaces face emotion. |
@@ -72,21 +72,46 @@ Source:
 | 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. |
-| s46 | S-46 | Final four lines / 3:58-4:16 | Muddy puddle with broken star reflection and boot edge. | C | anchor-test, expanded, priority | not started | Soul shot. Mud and stars both must be true. |
-| s47 | S-47 | Outro / 4:16-4:54 | Transition from muddy starlight to night road fire. | D | anchor-test, priority | not started | Tests D warmth after C. |
+| 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 | `shot-image-v1.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 | `shot-image-v1.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. |
-| s51 | S-51 | Outro / 4:16-4:54 | Final wide: night road, several far fires, tiny traveler. | D | anchor-test, expanded, priority | not started | Ending must remain unfinished and restrained. |
+| s51 | S-51 | Outro / 4:16-4:54 | Final wide: night road, several far fires, tiny traveler. | D | anchor-test, expanded, priority | accepted | `shot-image-v1.png` accepted by user. Ending remains unfinished and restrained. |
 
 ## Current Active Batch
 
 No prompt files exist yet.
 
-Next active batch after anchors:
+Five-shot trial plan/spec:
+
+- `five-shot-trial-plan-v1.md`
+
+Next active batch after plan approval:
 
 - `s01`
 - `s10`
 - `s46`
 - `s47`
 - `s51`
+
+Accepted shot:
+
+- `s01/shot-image-v1.png`
+- `s10/shot-image-v1.png`
+- `s46/shot-image-v1.png`
+- `s47/shot-image-v1.png`
+- `s51/shot-image-v1.png`
+
+Current shot brief:
+
+- `s51/shot-brief-v1.md`
+
+Current shot prompt/spec:
+
+- `s51/shot-prompt-v1.md`
+
+Current generated candidate:
+
+- `s51/shot-image-v1.png`
+- `s51/shot-review-v1.md`
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-board-v1.md b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-board-v1.md
new file mode 100644
index 0000000..d207dac
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-board-v1.md
@@ -0,0 +1,203 @@
+# Character Anchor Board V1
+
+## Status
+
+- Document status: pending review.
+- Execution status: non-generation-ready draft; do not generate from this file.
+- Compliance note: this draft was written before the repository rule requiring `gpt-image-2` Skill workflow for all project-bound image prompts. It must be rebuilt through the Skill, with mode/template/source records, before any image generation.
+- Target: one 16:9 character anchor board for the traveler in `在路上`.
+- Scope: main board only. Transparent character slices are a later step.
+
+## Purpose
+
+This board is a character consistency control sheet, not a finished MV shot, poster, portrait, or atmosphere image.
+
+It must lock:
+
+1. front / side / back body silhouette;
+2. hair, stubble, age, and weathered facial state;
+3. long coat, inner layer, old backpack, and old boots;
+4. walking side and walking back posture;
+5. the boundary between tired dignity and visual drift.
+
+Core question:
+
+> Later, when this person becomes small in a wide landscape shot, can we still recognize him as the same traveler?
+
+## Page Spec
+
+- Aspect ratio: 16:9 horizontal.
+- Background: very light neutral gray-white or very light warm gray.
+- Style: clean working character standard sheet, restrained documentary design.
+- Lighting: soft neutral studio-like light, no dramatic scene lighting.
+- Layout: one page with title area, character views, object breakdown, and short labels.
+- Image role: execution reference, not promotional art.
+
+## Board Layout
+
+### Top Title Area
+
+Title:
+
+> 《在路上》人物锚点板 / Character Anchor Board
+
+Subtitle:
+
+> 路上行者角色设定板
+
+One-line definition:
+
+> 50 岁左右的路上行者；沧桑摇滚，温暖疲惫，孤独但坚定。
+
+### Main View Area
+
+Use a 3-column working-board layout:
+
+1. left column: three-view core;
+2. middle column: state views;
+3. right column: clothing and prop breakdown.
+
+Required visual cells:
+
+| Cell | Content | Function |
+| --- | --- | --- |
+| A1 | front full body | age, body proportion, coat front, inner layer, overall temperament |
+| A2 | side full body | side silhouette, jaw/stubble profile, coat fall, backpack volume |
+| A3 | back full body | back-view recognition, rear hair shape, backpack, coat hem |
+| B1 | half-body state | face age, hair, stubble, tired dignity |
+| B2 | walking side | gait, coat movement, body weight, continuing-forward feeling |
+| B3 | walking back | key MV silhouette for later wide/back shots |
+| C1 | long coat / inner layer detail | color, length, worn texture, restrained shape |
+| C2 | old backpack detail | practical old backpack, used but not broken |
+| C3 | old boots detail | old walking boots, can carry mud, not fashion shoes |
+
+### Bottom Rule Area
+
+Keep the text short and functional.
+
+Keywords:
+
+- 50 岁左右
+- 路上行者
+- 沧桑摇滚
+- 温暖疲惫
+- 孤独但坚定
+- 被生活和系统磨过，但精神没塌
+
+Allowed:
+
+- wind-worn feeling
+- tiredness
+- practical old clothing
+- restrained ruggedness
+- mud and use traces
+
+Forbidden:
+
+- homeless coding
+- drunk coding
+- fashion model styling
+- heroic poster pose
+- polished travel blogger image
+- collapse, defeat, or visual pity
+
+## Character Definition
+
+The character is a 50-year-old male traveler.
+
+He is not a celebrity, not a fashion model, not a homeless man, not a drunk figure, and not a victorious hero. He is a person worn by life and systems, but still moving.
+
+### Appearance Anchors
+
+- dark, slightly messy hair;
+- visible stubble;
+- weathered but warm middle-aged face;
+- dark gray / dark khaki / coal-black long traveler's coat;
+- dark restrained inner layer;
+- old practical backpack;
+- old walking boots;
+- body capable of continuing forward, neither frail nor gym-styled.
+
+### Temperament Anchors
+
+- weathered rock feeling;
+- warm but tired;
+- lonely but steady;
+- silent, not defeated;
+- tired dignity rather than collapse.
+
+## Main Image Prompt
+
+Create a 16:9 horizontal character anchor board for a poetic road-documentary music video titled "在路上".
+
+The page is a clean working character standard sheet on a very light neutral gray-white background. It shows the same 50-year-old male traveler in multiple consistent views, with short functional labels and thin annotation lines. The design is restrained, practical, and cinematic, but not decorative.
+
+Character identity: a 50-year-old male traveler, weathered by life and systems, warm but tired, lonely but steady, silent and not defeated. Dark slightly messy hair, visible stubble, weathered middle-aged face, dark gray / dark khaki / coal-black long traveler's coat, dark inner layer, old practical backpack, old walking boots. The body should feel able to keep walking: neither frail nor athletic, neither stylish nor broken.
+
+Layout: top title area with "《在路上》人物锚点板 / Character Anchor Board" and the short definition "50 岁左右的路上行者；沧桑摇滚，温暖疲惫，孤独但坚定。"
+
+Main board content: three-view full-body core with front view, side view, and back view at consistent scale; state views with half-body state, walking side view, and walking back view; prop breakdowns for long coat and inner layer, old backpack, and old boots. Keep all views recognizable as the same person. Make the back view and walking back view especially clear because later MV shots will often show the traveler from behind or from far away.
+
+Board style: clean restrained film pre-production character sheet, light gray-white background, soft neutral light, readable silhouettes, low-saturation clothing colors, fine annotation lines, short labels only. The board should help later image generation keep the character consistent in wide landscapes, city edge shots, muddy road shots, and night fire shots.
+
+Emotional boundary: tired but dignified, worn but not ruined, lonely but still moving. The character should never become a portrait subject, heroic poster figure, fashion traveler, drunk man, or homeless visual stereotype.
+
+## Negative Prompt / Exclusion Rules
+
+Avoid:
+
+- front-facing glamour portrait;
+- heroic stance;
+- arms-open inspirational pose;
+- fashion editorial styling;
+- travel blogger image;
+- model-like grooming;
+- luxury outdoor gear;
+- cyberpunk, neon, sci-fi, UI, code, futuristic clothing;
+- homeless stereotype;
+- drunk stereotype;
+- dirty collapse;
+- exaggerated poverty;
+- comic-book character sheet style;
+- anime style;
+- glossy game concept art;
+- fantasy armor;
+- cinematic poster lighting;
+- dramatic landscape background;
+- muddy road scene background;
+- city background;
+- campfire background;
+- heavy gradient background;
+- colorful magazine layout;
+- text-heavy page.
+
+## Suggested Generation Settings
+
+These are generation guidance notes, not required tool commands.
+
+- Format: 16:9 horizontal.
+- Resolution target: high enough for text and small object labels to remain readable.
+- Background: light gray-white, nearly flat.
+- Character count: one character repeated across views, not multiple different people.
+- Text density: low. Use short labels only.
+- Priority: character consistency over beauty.
+- Most important cells: front full body, side full body, back full body, walking back.
+
+## Review Checklist
+
+Use this checklist before accepting the board.
+
+1. Do front, side, and back views clearly look like the same person?
+2. Is the back view recognizable without relying on the face?
+3. Does walking side / walking back feel real rather than posed?
+4. Is tiredness balanced with dignity?
+5. Are coat, backpack, and boots stable enough to guide later shots?
+6. Does the board avoid homeless, drunk, fashion model, hero poster, and travel blogger drift?
+7. When the character is mentally shrunk to a wide shot, do coat, backpack, boots, shoulder line, and gait still read as the same traveler?
+8. Is the board useful as an execution reference rather than just a good-looking illustration?
+
+## Pending User Review
+
+If approved, the next step is to generate the first character anchor board attempt from this document.
+
+Do not proceed to transparent slices, style anchor board, or five-shot tests until this document is accepted and the first board generation step is explicitly requested.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-brief-v1.md b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-brief-v1.md
new file mode 100644
index 0000000..75b99cf
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-brief-v1.md
@@ -0,0 +1,309 @@
+# Character Anchor Brief V1
+
+## Status
+
+- Document status: pending review.
+- Execution status: brief only; do not generate images from this file.
+- Source basis: primary character intake files plus approved trial plan.
+- Next gated action after approval: create `character-anchor-prompt-v1.md` through the `gpt-image-2` Skill workflow.
+
+## Purpose
+
+This brief defines the project-specific character anchor requirements for `在路上`.
+
+It is not a final image prompt. It is the execution source that the next Skill-based prompt document should use.
+
+The character anchor must answer:
+
+> When the traveler appears as a small side view, back view, or distant figure in later MV shots, can we still tell this is the same person?
+
+## Source Files Actually Used
+
+Primary sources:
+
+- `intake/阶段四-人物锚点板策划说明.md`
+- `intake/阶段四-人物锚点板页面结构建议稿.md`
+- `intake/阶段四-人物锚点板文案标注稿.md`
+
+Planning gate:
+
+- `visual-system/characters/character-anchor-trial-plan.md`
+
+Not used in this brief:
+
+- User source photo: not read yet. It will be handled in the prompt stage as an identity reference if the file is readable and approved for use.
+- `character-anchor-board-v1.md`: treated as a non-generation-ready draft, not as an authoritative source.
+
+## Source Photo Role
+
+The user source photo path is:
+
+```text
+C:\Users\wangq\Pictures\照片\photo.jpg
+```
+
+Role in later prompt work:
+
+- identity reference;
+- face direction and lived-person basis;
+- age, facial structure, hair / stubble direction if visible;
+- not an edit target;
+- not a passport-photo style target;
+- not a requirement to reproduce the exact original photo pose, lighting, background, or clothing.
+
+Expected transformation:
+
+> Use the source photo to keep a recognizable human basis, then translate the figure into the `在路上` traveler identity: weathered, restrained, road-worn, and usable across distant landscape-led shots.
+
+The prompt stage must explicitly label this input as `identity reference`.
+
+## Character Role In The MV
+
+The traveler is not the visual protagonist in a portrait-MV sense.
+
+He is a moving coordinate inside a larger world:
+
+- world before face;
+- road before pose;
+- silhouette before expression;
+- continuation before victory.
+
+The character anchor must support wide shots, back shots, side shots, muddy-road shots, city-edge shots, and night-fire shots.
+
+## Identity Definition
+
+### Core Identity
+
+- 50-year-old male traveler.
+- Based on the user's self-image direction, but translated into a road-film character.
+- Worn by life and systems, but still walking.
+- Not a celebrity, fashion model, homeless figure, drunk figure, or hero-poster protagonist.
+
+### Temperament
+
+- weathered rock feeling;
+- warm but tired;
+- lonely but steady;
+- silent, not defeated;
+- tired but not broken;
+- rugged but not collapsed;
+- road-worn dignity.
+
+### Body And Presence
+
+- body should feel capable of continuing forward;
+- neither frail nor gym-styled;
+- not an inspirational heroic body;
+- posture may be slightly weighted, but not defeated;
+- gait should be real, slow, and stable.
+
+## Appearance Anchors
+
+These features should remain stable across views and later shots.
+
+### Head / Face
+
+- dark, slightly messy hair;
+- visible stubble;
+- middle-aged weathering;
+- warm tiredness;
+- no polished grooming;
+- no fashion portrait expression;
+- no dramatic heroic stare.
+
+### Clothing
+
+- dark gray / dark khaki / coal-black long traveler's coat;
+- dark restrained inner layer;
+- coat has light use traces but is not torn or theatrical;
+- coat length and silhouette must remain stable enough for later back-view recognition;
+- clothing is practical and road-worn, not branded outdoor fashion.
+
+### Props
+
+- old practical backpack;
+- used but not broken;
+- medium volume, not oversized, not tactical-fashion, not luxury gear;
+- stable enough to identify the traveler from behind.
+
+### Boots
+
+- old walking boots;
+- can carry mud;
+- durable and road-appropriate;
+- not new fashion boots;
+- not shoe-ad closeup styling.
+
+## Required Main Board Scope
+
+The main anchor board should be one 16:9 horizontal page on a clean shallow gray-white or very light warm-gray background.
+
+The main board should prioritize identity consistency and execution usefulness over decorative beauty.
+
+### Required Visual Cells
+
+| ID | Cell | Purpose | Priority |
+| --- | --- | --- | --- |
+| A1 | Front full body | Establish age, body proportion, coat front, inner layer, and overall temperament. | Required |
+| A2 | Side full body | Establish side silhouette, jaw/stubble profile, coat fall, and backpack volume. | Required |
+| A3 | Back full body | Establish back-view recognition, backpack placement, rear hair outline, and coat hem. | Required |
+| B1 | Half-body state | Confirm face age, hair, stubble, eye mood, tired dignity. | Required |
+| B2 | Walking side | Confirm gait, coat movement, body weight, and forward motion. | Strongly recommended |
+| B3 | Walking back | Confirm the most important later MV silhouette. | Strongly recommended |
+| C1 | Coat / inner layer detail | Lock clothing system and color range. | Recommended |
+| C2 | Backpack detail | Lock the back-view identity prop. | Recommended |
+| C3 | Boots detail | Lock the road / mud / walking identity prop. | Recommended |
+
+### MVP Board If Generation Capacity Is Limited
+
+Minimum acceptable first board:
+
+1. front full body;
+2. side full body;
+3. back full body;
+4. half-body state;
+5. short boundary notes or empty annotation zones.
+
+If the model struggles with a full 9-cell board, the next repair strategy should be split generation rather than forcing all requirements into one overloaded image.
+
+## Layout Guidance
+
+Preferred structure:
+
+- top: title / board identity area;
+- left: three-view core;
+- middle: state views;
+- right: clothing and prop breakdown;
+- bottom: compact rule / boundary area.
+
+The board should look like a practical film pre-production reference sheet, not a colorful character poster.
+
+### Background
+
+Use:
+
+- shallow gray-white;
+- very light neutral gray;
+- very light warm gray.
+
+Avoid:
+
+- landscape background;
+- city background;
+- muddy road background;
+- campfire background;
+- strong gradient;
+- dramatic studio shadow;
+- promotional poster background.
+
+Reason:
+
+The background must not interfere with silhouette, coat color, hair shape, backpack form, or boot details.
+
+## Text And Label Strategy
+
+Labels matter, but generated readable text is not reliable enough to carry the source of truth.
+
+For this trial, use this split:
+
+### Image Model Responsibility
+
+The image may include:
+
+- simple visual zones;
+- small label-like blocks;
+- thin annotation lines;
+- empty callout zones;
+- minimal non-critical labels if the model can render them cleanly.
+
+The image must not be trusted as the only source for exact Chinese text.
+
+### Document Responsibility
+
+This brief and later prompt/review files carry the authoritative labels:
+
+- age;
+- temperament;
+- clothing anchors;
+- prop anchors;
+- allowed drift;
+- forbidden drift;
+- review checklist.
+
+### Later Layout Responsibility
+
+If a final readable board with exact Chinese labels is needed, add or repair labels in a later page/layout step rather than relying only on image generation.
+
+## Allowed Drift
+
+These variations are acceptable if identity remains stable:
+
+- mild mud or dust;
+- wind-worn texture;
+- slightly messy hair;
+- tired eyes;
+- used clothing traces;
+- restrained ruggedness;
+- small pose variation between views.
+
+## Forbidden Drift
+
+Avoid:
+
+- homeless stereotype;
+- drunk stereotype;
+- collapse or spiritual defeat;
+- fashion model styling;
+- polished travel blogger image;
+- luxury outdoor gear;
+- heroic poster pose;
+- arms-open inspirational posture;
+- celebrity portrait look;
+- glossy game concept art;
+- anime / cartoon styling;
+- fantasy armor or costume;
+- cyberpunk / neon / sci-fi clothing;
+- passport-photo style;
+- direct front-facing glamour portrait;
+- shoe advertisement framing;
+- background scene becoming more important than the character.
+
+## Review Questions
+
+Use these questions after the first generated board.
+
+1. Do front, side, and back views clearly look like the same person?
+2. Does the back view work without relying on the face?
+3. Does walking side / walking back look natural, not posed?
+4. Is the traveler tired but still dignified?
+5. Are coat, backpack, and boots stable enough to guide later shots?
+6. Does the board avoid homeless, drunk, fashion, hero, and travel-blogger drift?
+7. If the figure is mentally scaled down to 1/8 or 1/12 of the frame, does the silhouette still read?
+8. Is this board useful for later `s01/s10/s46/s47/s51` prompt work?
+
+## Trial-Specific Validation
+
+This brief should support the next prompt-stage validation:
+
+1. Can the `gpt-image-2` Skill template be adapted from generic character-sheet logic to a documentary road-film character board?
+2. Can a source photo be used as identity reference without turning the output into a portrait?
+3. Can the board preserve identity while keeping the character subordinate to future landscape-led composition?
+4. Can the model handle a 9-cell board, or should future workflow prefer split boards?
+5. Should exact labels be kept out of image generation and handled later?
+
+## Gated Next Step
+
+After this brief is approved, create:
+
+```text
+visual-system/characters/character-anchor-prompt-v1.md
+```
+
+That prompt document must:
+
+- run and record the `gpt-image-2` mode check;
+- read and record the relevant Skill template/reference files;
+- include this brief as a source;
+- include the source photo as an explicit identity reference if readable and approved;
+- separate image-model output from text annotation responsibilities;
+- wait for user approval before generation.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-prompt-v1.md b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-prompt-v1.md
new file mode 100644
index 0000000..780477c
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-prompt-v1.md
@@ -0,0 +1,313 @@
+# Character Anchor Prompt V1
+
+## Status
+
+- Document status: pending review.
+- Execution status: prompt only; do not generate images before user approval.
+- Task type: reference-guided character anchor board generation.
+- Project: `在路上`.
+- User adjustment: add height `180cm`; make hair longer, beard longer, and beard denser.
+
+## Skill Mode Check
+
+Command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+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)"
+}
+```
+
+Execution interpretation:
+
+- Garden local generation is not enabled.
+- No `OPENAI_API_KEY` is available for Mode A.
+- This document uses the `gpt-image-2` Skill as a structured prompt-engineering workflow.
+- If image generation is later approved in this Codex environment, use the host image tool path as Mode B.
+- If no host image tool is used, this prompt remains an Advisor prompt for external/manual execution.
+
+## Skill References 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\portraits-and-characters\character-sheet.md`
+
+Template adaptation:
+
+- Base template family: `portraits-and-characters/character-sheet.md`.
+- Adaptation required: the base template targets animation/game/IP character sheets; this project requires a realistic documentary road-film character reference sheet.
+- Text policy adaptation: do not rely on generated readable labels as the source of truth.
+
+## Project Sources Used
+
+- `visual-system/characters/character-anchor-trial-plan.md`
+- `visual-system/characters/character-anchor-brief-v1.md`
+- `intake/阶段四-人物锚点板策划说明.md`
+- `intake/阶段四-人物锚点板页面结构建议稿.md`
+- `intake/阶段四-人物锚点板文案标注稿.md`
+
+Not used as authoritative source:
+
+- `visual-system/characters/character-anchor-board-v1.md`, because it is a non-generation-ready direct draft.
+
+## Input Image Role
+
+Declared source photo:
+
+```text
+C:\Users\wangq\Pictures\照片\photo.jpg
+```
+
+Input role:
+
+- identity reference;
+- face direction and lived-person basis;
+- not an edit target;
+- not a passport-photo style target;
+- not a requirement to preserve the original photo background, pose, lighting, or clothing.
+
+Readability status:
+
+- not loaded in this document-only step;
+- must be loaded or made accessible before generation;
+- if the host image tool cannot read the file from this path, copy or approve access before generation.
+
+Declared satisfactory sample:
+
+```text
+visual-system/characters/character-chatgpt.png
+```
+
+Sample role:
+
+- visual target reference for board structure and character-asset direction;
+- use for the clean shallow-gray reference-sheet layout, multi-view organization, dark long coat, old backpack, old boots, restrained working-board tone, and overall satisfactory direction;
+- not the primary identity reference;
+- not a requirement to copy every facial detail exactly;
+- not a final accepted asset until reviewed in this workflow.
+
+## Prompt Strategy
+
+Use one generated board to test whether a single image can stabilize the traveler identity.
+
+Primary output:
+
+- one 16:9 horizontal character anchor board;
+- clean shallow gray-white / very light warm-gray background;
+- same traveler shown across multiple views;
+- realistic film pre-production reference style;
+- minimal label zones, not text-heavy layout.
+
+Do not ask the image model to render exact Chinese labels as the source of truth. Exact labels and review criteria remain in project documents.
+
+If the full 9-cell board fails, the next iteration should split the work instead of forcing all requirements into one overloaded prompt.
+
+## Rendered Prompt
+
+Use the attached source photo as Input Image 1. Treat Input Image 1 only as an identity reference for the traveler's face direction, lived-person basis, age impression, hair/stubble direction, and facial structure. Do not copy the original photo background, pose, lighting, clothing, or passport-photo style.
+
+Use the attached `character-chatgpt.png` sample as Input Image 2 if available. Treat Input Image 2 as a visual target reference for the clean shallow-gray character-sheet layout, multi-view organization, dark long coat, old backpack, old boots, working-board tone, and overall satisfactory direction. Do not treat Input Image 2 as the primary identity source.
+
+Generate one 16:9 horizontal realistic character anchor board for the music-video project "在路上". The board is a practical film pre-production reference sheet for a landscape-led road documentary / poetic realism MV. It is not a finished MV shot, not a poster, not a portrait, and not promotional art.
+
+The character is a 50-year-old male traveler, about 180cm tall, translated from the identity reference into a road-film figure: weathered by life and systems, warm but tired, lonely but steady, silent and not defeated. He should feel capable of continuing forward. He is not frail, not gym-styled, not heroic, not fashionable, not homeless-coded, and not drunk-coded.
+
+Overall visual style: realistic documentary character reference, restrained cinematic realism, practical working board, clean shallow gray-white or very light warm-gray background, soft neutral lighting, readable silhouettes, low saturation, no dramatic scene lighting, no landscape or city background. The board should help later image generation keep the same traveler recognizable in wide road shots, back views, side views, muddy-road shots, city-edge shots, and night-fire shots.
+
+Character fixed attributes:
+
+- dark, slightly messy, somewhat longer hair;
+- longer and denser beard/stubble, visibly fuller than light stubble but still realistic and restrained;
+- weathered middle-aged face with tired dignity;
+- dark gray / dark khaki / coal-black long traveler's coat;
+- dark restrained inner layer;
+- old practical backpack, medium volume, used but not broken;
+- old walking boots, durable and road-worn, able to carry mud;
+- body posture slightly weighted but not defeated;
+- gait slow, real, and stable.
+
+Board layout:
+
+- top area: a quiet board identity area, with minimal title-like visual treatment but no reliance on readable text;
+- left area: three-view core with front full body, side full body, and back full body at consistent scale;
+- middle area: state views with half-body state, walking side view, and walking back view;
+- right area: clothing and prop detail cells for long coat / inner layer, old backpack, and old boots;
+- bottom area: compact visual rule strip or empty annotation zone for later labels.
+
+Cell requirements:
+
+1. Front full body: establish age, body proportion, coat front, inner layer, and overall temperament.
+2. Side full body: establish side silhouette, jaw/stubble profile, coat fall, and backpack volume.
+3. Back full body: establish rear hair outline, backpack placement, coat hem, and back-view recognition.
+4. Half-body state: confirm hair, stubble, tired eyes, and dignified middle-aged presence.
+5. Walking side: show real forward motion, coat movement, shoulder weight, and stable gait.
+6. Walking back: emphasize the later MV silhouette; back view must still feel like the same traveler.
+7. Coat / inner layer detail: show dark, practical, road-worn clothing, not fashion styling.
+8. Backpack detail: old practical backpack, used but not broken, not tactical fashion, not oversized.
+9. Boots detail: old walking boots, durable, slightly muddy or worn, not new fashion shoes and not shoe-ad framing.
+
+Text and label handling:
+
+- The image may include thin callout lines, simple label blocks, empty annotation zones, or simple alphanumeric badges.
+- Do not render long readable Chinese or English labels.
+- Do not make the image depend on exact text correctness.
+- Leave enough clean space for later document/page labels if needed.
+
+Composition priorities:
+
+- same-person consistency across all views;
+- back-view recognition without the face;
+- stable coat / backpack / boots system;
+- clear silhouette when mentally scaled down;
+- restrained working-board design;
+- character identity over decorative beauty.
+
+Avoid all of the following:
+
+- front-facing glamour portrait;
+- passport photo style;
+- heroic poster stance;
+- arms-open inspirational pose;
+- fashion editorial styling;
+- polished travel blogger image;
+- luxury outdoor gear;
+- homeless stereotype;
+- drunk stereotype;
+- collapse, pity, or spiritual defeat;
+- celebrity portrait look;
+- glossy game concept art;
+- anime / cartoon styling;
+- fantasy armor or costume;
+- cyberpunk, neon, sci-fi, UI, code, futuristic clothing;
+- muddy road scene background;
+- city background;
+- campfire background;
+- strong gradient background;
+- dramatic studio shadow;
+- shoe advertisement framing;
+- text-heavy page;
+- exact readable Chinese text as a critical output requirement;
+- multiple different men instead of one consistent traveler.
+
+The result should feel like a usable character standard sheet for the road-film traveler: a person whose face, hair, stubble, coat, backpack, boots, posture, and walking silhouette can guide later landscape-led MV images while keeping the world larger than the person.
+
+## Structured Prompt Map
+
+```json
+{
+  "type": "realistic_documentary_character_anchor_board",
+  "goal": "Generate a 16:9 character anchor board that stabilizes the same traveler across front, side, back, state, walking, clothing, backpack, and boot references for later MV shot generation.",
+  "input_images": [
+    {
+      "id": "Input Image 1",
+      "path": "C:\\Users\\wangq\\Pictures\\照片\\photo.jpg",
+      "role": "identity reference",
+      "use": "face direction, lived-person basis, age impression, longer hair / denser beard direction, facial structure",
+      "do_not_use_for": "passport-photo style, exact pose, original background, original lighting, original clothing"
+    },
+    {
+      "id": "Input Image 2",
+      "path": "projects/2026-06-23-在路上/visual-system/characters/character-chatgpt.png",
+      "role": "visual target reference",
+      "use": "clean shallow-gray character-sheet layout, multi-view organization, dark long coat, old backpack, old boots, restrained working-board tone, satisfactory direction",
+      "do_not_use_for": "primary identity source, exact face copy, final acceptance without review"
+    }
+  ],
+  "character": {
+    "identity": "50-year-old male traveler",
+    "height": "about 180cm",
+    "temperament": "weathered by life and systems, warm but tired, lonely but steady, silent and not defeated",
+    "hair": "dark, slightly messy, somewhat longer",
+    "face": "longer and denser beard/stubble, visibly fuller than light stubble, weathered middle-aged face, tired dignity",
+    "clothing": "dark gray / dark khaki / coal-black long traveler's coat, dark restrained inner layer",
+    "props": "old practical backpack, medium volume, used but not broken",
+    "boots": "old walking boots, durable, road-worn, can carry mud",
+    "body_presence": "capable of continuing forward, neither frail nor gym-styled"
+  },
+  "layout": {
+    "aspect_ratio": "16:9 horizontal",
+    "background": "clean shallow gray-white or very light warm-gray",
+    "style": "practical film pre-production character reference sheet",
+    "sections": [
+      "front full body",
+      "side full body",
+      "back full body",
+      "half-body state",
+      "walking side",
+      "walking back",
+      "coat / inner layer detail",
+      "backpack detail",
+      "boots detail"
+    ],
+    "text_policy": "minimal label-like zones only; no long readable Chinese or English labels; exact labels belong in project documents"
+  },
+  "constraints": {
+    "must_keep": [
+      "same person across all views",
+      "back-view recognition without face",
+      "stable long coat / backpack / boots system",
+      "tired dignity rather than collapse",
+      "usable silhouette for later wide shots"
+    ],
+    "avoid": [
+      "portrait MV framing",
+      "fashion model styling",
+      "homeless or drunk coding",
+      "hero poster pose",
+      "travel blogger image",
+      "passport photo style",
+      "text-heavy board",
+      "exact readable text dependency",
+      "multiple different men"
+    ]
+  }
+}
+```
+
+## Acceptance Checklist
+
+Review after generation:
+
+1. Front, side, and back views look like the same person.
+2. Back view is recognizable without relying on the face.
+3. Walking side and walking back are natural, not posed.
+4. The figure is tired but dignified, not collapsed.
+5. Coat, backpack, and boots are stable enough for later shot prompts.
+6. The board avoids homeless, drunk, fashion, hero, and travel-blogger drift.
+7. The silhouette would remain identifiable at 1/8 or 1/12 frame scale.
+8. The image does not depend on exact generated text.
+9. The board is useful for later `s01/s10/s46/s47/s51` prompt work.
+
+## Repair Strategy
+
+If the first generation fails:
+
+- If views look like different people: reduce scope to front / side / back / half-body MVP and regenerate.
+- If labels are garbled: remove all readable text requirements and use empty label zones only.
+- If the source photo dominates too much: reduce portrait fidelity and emphasize road-film traveler translation.
+- If the figure becomes fashionable or heroic: strengthen clothing/proportion/pose negatives.
+- If the figure becomes homeless or broken: strengthen dignity and practical-road-worn constraints.
+- If nine cells are crowded: split into main board and later prop/detail boards.
+
+## Gated Next Step
+
+Do not generate from this prompt until the user approves this document.
+
+When approved, the generation step must:
+
+1. make `C:\Users\wangq\Pictures\照片\photo.jpg` available to the host image tool as Input Image 1;
+2. pass the rendered prompt above to the active image-generation path;
+3. save the generated output into the project, not only in a tool default location;
+4. create a review record before any transparent character slices.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-prompt-v2.md b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-prompt-v2.md
new file mode 100644
index 0000000..9efb231
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-prompt-v2.md
@@ -0,0 +1,163 @@
+# Character Anchor Prompt V2
+
+## Status
+
+- Document status: generation test prompt.
+- Execution status: used for one Codex host generation test.
+- Revision focus: preserve the preferred character image from v1, but correct the walking posture.
+- Project: `在路上`.
+
+## Skill Mode Check
+
+Command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+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)"
+}
+```
+
+Interpretation:
+
+- Use `gpt-image-2` Skill as prompt-engineering workflow.
+- Use Codex host image generation path for this test.
+- Do not call Garden scripts.
+
+## Revision Inputs
+
+Primary identity reference:
+
+- `C:\Users\wangq\Pictures\照片\photo.jpg`
+
+Preferred character-image reference:
+
+- `visual-system/characters/character-anchor-v1-codex-test.png`
+
+Walking-posture reference:
+
+- `visual-system/characters/character-chatgpt.png`
+
+User feedback:
+
+- Prefer the character image in `character-anchor-v1-codex-test.png`.
+- Do not keep the hunched walking pose from `character-anchor-v1-codex-test.png`.
+- Walking posture should be closer to `character-chatgpt.png`: upright chest, straighter upper body, not bent at the waist.
+
+## Rendered Prompt
+
+Use the red-background passport-style headshot visible in this conversation as the primary identity reference. Preserve the face direction, lived-person basis, age impression, facial structure, glasses, and general likeness direction, but do not copy the red background, suit, passport-photo pose, lighting, or clothing.
+
+Use `character-anchor-v1-codex-test.png` as the preferred character-image reference. Preserve its overall traveler identity direction: longer messy dark hair, fuller beard, weathered middle-aged face, dark long road-worn coat, old backpack, old boots, shallow gray-white character-board layout, practical documentary reference-sheet tone. This is the preferred character look.
+
+Use `character-chatgpt.png` as the positive walking-posture reference. In walking side and walking back cells, the traveler should walk with an upright chest, a straighter spine, natural head and neck alignment, relaxed shoulders that do not collapse forward, and a forward-moving body that is steady rather than hunched. The walk should feel like a tired but dignified traveler who still carries himself, not a person bent under defeat.
+
+Generate one 16:9 horizontal realistic character anchor board for the music-video project "在路上". The board is a practical film pre-production reference sheet for a landscape-led road documentary / poetic realism MV. It is not a finished MV shot, not a poster, not a portrait, and not promotional art.
+
+The character is a 50-year-old male traveler, about 180cm tall, translated from the identity reference into a road-film figure: weathered by life and systems, warm but tired, lonely but steady, silent and not defeated. He should feel capable of continuing forward. He is not frail, not gym-styled, not heroic, not fashionable, not homeless-coded, and not drunk-coded.
+
+Character fixed attributes:
+
+- dark, slightly messy, somewhat longer hair;
+- longer and denser beard/stubble, visibly fuller than light stubble but still realistic and restrained;
+- weathered middle-aged face with tired dignity;
+- dark gray / dark khaki / coal-black long traveler's coat;
+- dark restrained inner layer;
+- old practical backpack, medium volume, used but not broken;
+- old walking boots, durable and road-worn, able to carry mud;
+- body posture solid and capable of continuing forward;
+- walking posture upright and steady, not hunched.
+
+Board layout:
+
+- clean shallow gray-white or very light warm-gray background;
+- soft neutral lighting;
+- low-saturation realistic documentary character reference;
+- top area as a quiet board identity zone with minimal title-like visual treatment but no reliance on readable text;
+- left area with three-view core: front full body, side full body, back full body at consistent scale;
+- middle area with state views: half-body state, walking side view, walking back view;
+- right area with clothing and prop detail cells for long coat / inner layer, old backpack, and old boots;
+- bottom area as compact visual rule strip or empty annotation zone for later labels.
+
+Cell requirements:
+
+1. Front full body: preserve the preferred v1 character image direction, with solid adult body proportion and road-worn dignity.
+2. Side full body: preserve side silhouette and coat/backpack system without making the torso collapse forward.
+3. Back full body: preserve back-view recognition, backpack placement, rear hair outline, and coat hem.
+4. Half-body state: preserve the preferred v1 face direction, longer hair, fuller beard, tired eyes, and dignified middle-aged presence.
+5. Walking side: must show an upright walking posture: chest open, spine straighter, head naturally lifted, shoulders relaxed but not rounded forward. Do not bend the waist. Do not hunch the back. Do not make the traveler look defeated or physically collapsed.
+6. Walking back: must show a steady upright back view while walking, with shoulders not folded forward and the backpack sitting naturally.
+7. Detail cells: show dark practical road-worn coat, old practical backpack, and old durable boots.
+
+Text and label handling:
+
+- The image may include thin callout lines, simple label blocks, empty annotation zones, or simple alphanumeric badges.
+- Do not render long readable Chinese or English labels.
+- Do not make the image depend on exact text correctness.
+- Leave enough clean space for later document/page labels if needed.
+
+Composition priorities:
+
+- preserve the v1 preferred character image;
+- correct the walking posture to be upright like the ChatGPT sample;
+- same-person consistency across all views;
+- back-view recognition without face;
+- stable coat / backpack / boots system;
+- clear silhouette when mentally scaled down;
+- restrained working-board design;
+- character identity over decorative beauty.
+
+Strictly avoid:
+
+- hunched walking posture;
+- bent waist in walking side view;
+- rounded shoulders collapsing forward;
+- bowed head / defeated neck posture;
+- stooped, frail, elderly-collapse walk;
+- front-facing glamour portrait;
+- passport photo style;
+- heroic poster stance;
+- arms-open inspirational pose;
+- fashion editorial styling;
+- polished travel blogger image;
+- luxury outdoor gear;
+- homeless stereotype;
+- drunk stereotype;
+- collapse, pity, or spiritual defeat;
+- celebrity portrait look;
+- glossy game concept art;
+- anime/cartoon styling;
+- fantasy armor or costume;
+- cyberpunk/neon/sci-fi/UI/code/futuristic clothing;
+- muddy road scene background;
+- city background;
+- campfire background;
+- strong gradient background;
+- dramatic studio shadow;
+- shoe advertisement framing;
+- text-heavy page;
+- exact readable Chinese text as a critical output requirement;
+- multiple different men instead of one consistent traveler.
+
+The result should keep the preferred v1 traveler image while replacing the hunched walking pose with an upright, chest-open, dignified walking posture similar to the ChatGPT sample.
+
+## Acceptance Delta From V1
+
+V2 passes only if:
+
+1. The character image remains closer to `character-anchor-v1-codex-test.png` than to a new unrelated person.
+2. The walking side view is visibly less hunched than v1.
+3. The walking side view has an upright chest and a straighter upper body.
+4. The walking back view does not read as defeated or collapsed.
+5. Coat, backpack, boots, hair, and beard remain stable.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-review-v1.md b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-review-v1.md
new file mode 100644
index 0000000..3d2cf20
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-review-v1.md
@@ -0,0 +1,105 @@
+# Character Anchor Review V1
+
+## Status
+
+- Review status: superseded by v2 review.
+- Image acceptance state: not final anchor; v2 was accepted later.
+- Generation purpose: test Codex host image generation capability and character-anchor workflow completeness.
+- Next gated action: completed by `character-anchor-review-v2.md` and `character-identity-lock.md`.
+
+## Generation Context
+
+Prompt document:
+
+- `visual-system/characters/character-anchor-prompt-v1.md`
+
+Generated image:
+
+- `visual-system/characters/character-anchor-v1-codex-test.png`
+
+Tool default source retained:
+
+- `C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_098d4e68be1207f7016a3a2cde9678819889913b70ad004b89.png`
+
+Image properties:
+
+- Width: 1672
+- Height: 941
+- Pixel format: RGB
+- Aspect: close to 16:9
+
+## Inputs Used
+
+### Identity Reference
+
+```text
+C:\Users\wangq\Pictures\照片\photo.jpg
+```
+
+Role:
+
+- identity reference;
+- not edit target;
+- not passport-photo style target.
+
+### Satisfactory Sample Reference
+
+```text
+visual-system/characters/character-chatgpt.png
+```
+
+Role:
+
+- visual target reference for board structure and satisfactory character-asset direction;
+- not the primary identity source.
+
+### User Prompt Adjustments
+
+- Add height: about 180cm.
+- Hair and beard should be longer.
+- Beard should be denser / fuller.
+
+## Privacy Note
+
+External MiniMax visual analysis was attempted for `character-chatgpt.png` but rejected by the environment because sending a project-local image to an external service would be private workspace data exfiltration. The review below is based on local visual inspection only.
+
+## Review Checklist
+
+| Check | Result | Notes |
+| --- | --- | --- |
+| Front / side / back same-person consistency | Pass | The front, side, and back views read as the same traveler. |
+| Back-view recognition without face | Pass | Long coat, backpack, hair shape, and stance are consistent enough for later back-view use. |
+| Walking side / walking back naturalness | Pass | Walking side and rear walking pose are natural and not heroic. |
+| Tired but dignified | Pass | Face and posture read as restrained, weathered, and dignified rather than collapsed. |
+| 180cm body direction | Pass with note | The body reads adult, solid, and tall enough; exact height cannot be verified from the board alone. |
+| Longer hair / denser beard | Pass | Compared with the prior sample, hair and beard are clearly longer and fuller. |
+| Coat / backpack / boots stability | Pass | The old long coat, backpack, and boots are strong and reusable visual anchors. |
+| Avoid homeless / drunk / fashion / hero drift | Pass | It stays practical and road-worn without becoming a fashion shoot or poverty stereotype. |
+| Avoid text dependence | Pass | The board uses minimal blocks and empty label areas; exact labels are not required from the image. |
+| Usefulness for later five-shot prompts | Pass | Good enough to test character carryover into `s01/s10/s46/s47/s51`. |
+
+## Observed Strengths
+
+- The generated board successfully follows the multi-view character-sheet structure.
+- The shallow gray-white background keeps silhouette and clothing readable.
+- The dark long coat, old backpack, and old boots are stronger than the face as wide-shot identity anchors.
+- The longer hair and fuller beard move the character closer to the requested road-worn traveler image.
+- The side and back views are useful for later landscape-led composition.
+
+## Observed Risks
+
+- The result strongly follows the ChatGPT sample layout, which is good for this test but may make future iterations too dependent on one sample.
+- The face is directionally related to the source photo, but not a strict likeness match. This may be acceptable because the brief asks for a translated road-film figure, not a portrait copy.
+- The board is strong, but still not proof that later wide environmental shots will preserve identity; that must be tested in the five-shot package.
+- Exact 180cm cannot be visually audited without scale marks or a comparison object.
+
+## Recommendation
+
+This is a successful process test and a plausible first character anchor candidate.
+
+Recommended next decision:
+
+1. If the user accepts this board as the first character anchor candidate, create `character-identity-lock.md`.
+2. If the user wants a tighter likeness to the source photo, regenerate with stronger identity-reference wording.
+3. If the user wants less dependency on the ChatGPT sample, regenerate with the same identity but lighter sample-reference wording.
+4. Do not start transparent slices until this board is accepted or a repaired version replaces it.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-review-v2.md b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-review-v2.md
new file mode 100644
index 0000000..86cfc87
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-review-v2.md
@@ -0,0 +1,71 @@
+# Character Anchor Review V2
+
+## Status
+
+- Review status: user selected v2.
+- Image acceptance state: accepted as the current main character anchor board.
+- Generation purpose: repair walking posture while preserving the preferred v1 character image.
+- Next gated action: review `character-identity-lock.md` before any transparent slices or scene prompts.
+
+## Generation Context
+
+Prompt document:
+
+- `visual-system/characters/character-anchor-prompt-v2.md`
+
+Generated image:
+
+- `visual-system/characters/character-anchor-v2-codex-posture-test.png`
+
+Tool default source retained:
+
+- `C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_059d9628223dbafb016a3a2f45e45c8198a75f641eb5f2295f.png`
+
+Image properties:
+
+- Width: 1672
+- Height: 941
+- Pixel format: RGB
+- Aspect: close to 16:9
+
+## User Feedback Addressed
+
+Original feedback:
+
+- Character image in `character-anchor-v1-codex-test.png` is preferred.
+- Walking posture in v1 is too hunched.
+- Regenerate with walking posture closer to `character-chatgpt.png`: upright chest, not bent at the waist.
+
+V2 prompt response:
+
+- Preserved v1 as preferred character-image reference.
+- Used `character-chatgpt.png` as positive walking-posture reference.
+- Added strict negative constraints against hunched walking, bent waist, rounded shoulders, bowed head, and collapsed posture.
+
+## Review Checklist
+
+| Check | Result | Notes |
+| --- | --- | --- |
+| Preserve v1 character image direction | Pass | Face, hair, beard, coat, backpack, and boots remain close to v1. |
+| Walking side less hunched than v1 | Pass | The main central walking side cell is visibly more upright. |
+| Chest / upper body more open | Pass | Main walking pose has a straighter upper body and less collapsed shoulder line. |
+| Walking back not defeated | Pass | Rear walking pose is steady and does not read as collapsed. |
+| Same-person consistency | Pass | Multi-view identity remains coherent. |
+| Coat / backpack / boots stability | Pass | Object system remains strong and consistent. |
+| Text dependence avoided | Pass | The board does not rely on exact generated labels. |
+
+## Remaining Notes
+
+- The small lower-left walking pose still has mild forward lean, but it reads as motion rather than waist-bent defeat.
+- The main walking side cell is the strongest posture repair and is closer to the desired "挺胸走" direction.
+- Compared with v1, v2 has slightly more structured layout and a somewhat less intimate half-body portrait, but the identity direction remains usable.
+
+## Recommendation
+
+V2 is a better candidate than v1 if walking posture is the priority.
+
+Suggested decision:
+
+1. If the user accepts v2, use `character-anchor-v2-codex-posture-test.png` as the first accepted character anchor candidate and create `character-identity-lock.md`.
+2. If the user wants one more repair, keep v2 as the base and further specify: "all walking cells must be upright; no secondary walking pose may lean forward."
+3. Do not create transparent slices until a character anchor candidate is accepted.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-trial-plan.md b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-trial-plan.md
new file mode 100644
index 0000000..8f70aec
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-trial-plan.md
@@ -0,0 +1,313 @@
+# Character Anchor Trial Plan
+
+## Status
+
+- Document status: pending review.
+- Execution status: planning only; do not generate images from this file.
+- Scope: character anchor trial for `在路上`.
+- Next gated action after approval: rebuild the character anchor prompt through the `gpt-image-2` Skill workflow.
+
+## Why This Plan Exists
+
+The current character work has two separate goals:
+
+1. deliver the character anchor assets needed by this MV project;
+2. validate which parts of the character-anchor workflow should become reusable Video Workbench knowledge later.
+
+These goals must stay separate. Project-specific execution should not be mistaken for public workflow rules, and public workflow candidates should not be promoted before they have been tested through real image generation and review.
+
+## Current Issue
+
+`character-anchor-board-v1.md` is a useful content draft, but it is not generation-ready.
+
+Problems:
+
+1. It was drafted directly from project summaries and intake notes instead of the `gpt-image-2` Skill workflow.
+2. It does not record Skill mode, template/reference path, or source files actually used.
+3. It does not clearly separate image-model responsibilities from later document/page annotation responsibilities.
+4. It does not include the user's source photo as an explicit identity reference input.
+5. It compresses character-board, transparent-slice, and later execution-asset work into one document.
+
+Therefore, it should be treated only as a non-generation-ready draft and rebuilt before any image generation.
+
+## Source Reliability Assessment
+
+### Primary Sources
+
+These are the most reliable project sources for character-anchor work:
+
+| Source | Role | Reliability |
+| --- | --- | --- |
+| `intake/阶段四-人物锚点板策划说明.md` | Defines why a character board is needed, background strategy, output types, and review priorities. | High as accepted GPT planning; still requires Codex execution judgment. |
+| `intake/阶段四-人物锚点板页面结构建议稿.md` | Defines page structure, layout modules, labels, and review focus. | High for board structure; must be adapted for image-model limits. |
+| `intake/阶段四-人物锚点板文案标注稿.md` | Explains why labels matter and which identity features should be stabilized. | High for review logic; image text must be handled carefully. |
+| User source photo, once provided/readable | Identity reference for face direction and lived-person basis. | High for identity direction; not an edit target and not a passport-photo style target. |
+
+### Execution Control Sources
+
+These files are useful but secondary:
+
+| Source | Role | Reliability |
+| --- | --- | --- |
+| `visual-system/visual-system.md` | Condensed accepted visual system and materialization targets. | Reliable summary, but not a substitute for primary intake. |
+| `execution-plan.md` | Current phase, batch order, and gate sequence. | Reliable for sequencing, not for detailed prompt construction. |
+| `project.md` | Project state and next action. | Reliable for status only. |
+
+### Skill / Workflow Sources
+
+These govern prompt production:
+
+| Source | Role | Reliability |
+| --- | --- | --- |
+| `gpt-image-2` Skill `SKILL.md` | Required mode-aware workflow for image prompts and generation. | Required workflow source. |
+| `gpt-image-2/references/prompt-writing.md` | Template design method: fields, source mapping, parameter strategy. | Required when building structured prompt documents. |
+| `gpt-image-2/references/portraits-and-characters/character-sheet.md` | Closest current template for character board / three-view sheet. | Useful but not perfect; must be adapted from anime/IP defaults to documentary road-film character reference. |
+| `investigations/2026-06-23-gpt-image-2-advisor-prompt-quality-v2-broader-use-cases/提示词质量综合调研报告.md` | Evidence that Advisor prompts outperform direct prompts for Video Workbench prompt quality. | Strong process evidence; final image quality still unverified. |
+
+## Assessment Of GPT Character Drafts
+
+The three GPT character files are directionally sound but should not be treated as final execution.
+
+### What Looks Reasonable
+
+- Use a shallow gray-white or light warm-gray clean background for the main board.
+- Create a board for identity consistency, not a finished MV shot.
+- Prioritize front / side / back full-body views.
+- Include half-body state, walking side, and walking back to test motion identity.
+- Track old coat, old backpack, and old boots as core recognition anchors.
+- Review the balance between tiredness and dignity.
+- Avoid fashion model, homeless, drunk, hero-poster, and travel-blogger drift.
+
+### What Needs Adaptation
+
+- GPT assumes text labels can be part of the board; image models may create unreliable or fake text.
+- GPT mixes main board and transparent slices; these should be separate gated steps.
+- GPT gives a human-readable board plan, not a `gpt-image-2` Advisor prompt.
+- GPT does not define the source photo role.
+- GPT does not separate project-specific identity from reusable character-anchor workflow.
+
+## Should We Extract The GPT Drafts?
+
+Yes, but into two different layers.
+
+### Layer 1: Project Execution Extraction
+
+Extract what this project needs:
+
+- traveler identity;
+- board layout;
+- view list;
+- clothing / prop anchors;
+- allowed and forbidden drift;
+- source-photo role;
+- review checklist;
+- generation gate sequence.
+
+This becomes the basis for the rebuilt project prompt.
+
+### Layer 2: Public Workflow Candidate Extraction
+
+Extract what might become reusable later:
+
+- character anchor purpose definition;
+- source hierarchy;
+- standard deliverables;
+- prompt workflow;
+- text-label handling rule;
+- review rubric;
+- gate sequence.
+
+This layer remains a candidate until validated by this trial.
+
+## Project-Specific Character Work Package
+
+The current project should perform the following work, one gated step at a time.
+
+### Step 1: Trial Plan
+
+Output:
+
+- `visual-system/characters/character-anchor-trial-plan.md`
+
+Purpose:
+
+- decide what is being validated;
+- separate project work from reusable workflow candidates;
+- identify source reliability and prompt-production rules.
+
+Gate:
+
+- user review.
+
+### Step 2: Character Execution Brief
+
+Output:
+
+- `visual-system/characters/character-anchor-brief-v1.md`
+
+Purpose:
+
+- extract the accepted character definition from primary intake sources;
+- record source-photo role;
+- define required views and objects;
+- define review criteria before prompt writing.
+
+Gate:
+
+- user review.
+
+### Step 3: Skill-Based Prompt Document
+
+Output:
+
+- `visual-system/characters/character-anchor-prompt-v1.md`
+
+Purpose:
+
+- run and record the `gpt-image-2` mode check;
+- record template/reference files used;
+- map source material into a structured Advisor prompt;
+- separate image-model output from later annotation/document responsibilities;
+- include negative constraints and acceptance criteria.
+
+Gate:
+
+- user review before generation.
+
+### Step 4: Main Character Anchor Board Generation
+
+Output:
+
+- first generated main board image;
+- prompt trace if required by the active Skill mode;
+- `visual-system/characters/character-anchor-review-v1.md`
+
+Purpose:
+
+- test whether a single board can lock the character consistently enough.
+
+Gate:
+
+- user review before any transparent slices.
+
+### Step 5: Transparent Character Slices
+
+Output candidates:
+
+- front full body;
+- side full body;
+- back full body;
+- walking side;
+- walking back.
+
+Purpose:
+
+- create execution references for later shot prompts or compositing reference.
+
+Gate:
+
+- only after main board is accepted.
+
+### Step 6: Character Identity Lock
+
+Output:
+
+- `visual-system/characters/character-identity-lock.md`
+
+Purpose:
+
+- record the accepted stable identity features after image review;
+- provide a compact reference for style anchor and five-shot prompt work.
+
+Gate:
+
+- only after at least one board or slice set is accepted.
+
+## Candidate Reusable Character-Anchor Structure
+
+These items are candidates for future Video Workbench knowledge, but should not be promoted until this project validates them.
+
+### Candidate Standard Deliverables
+
+1. `character-anchor-trial-plan.md`
+2. `character-anchor-brief-v1.md`
+3. `character-anchor-prompt-v1.md`
+4. generated main board image
+5. `character-anchor-review-v1.md`
+6. transparent slices, when needed
+7. `character-identity-lock.md`
+
+### Candidate Source Hierarchy
+
+1. user-provided source/reference images;
+2. accepted GPT intake files;
+3. project-local visual-system summary;
+4. project-local execution plan;
+5. prior generated assets and review records.
+
+### Candidate Review Rubric
+
+1. same-person consistency across front / side / back;
+2. back-view recognition without face;
+3. walking state naturalness;
+4. identity stability when scaled down;
+5. clothing and prop stability;
+6. balance between expressive character and non-portrait role;
+7. project-specific drift avoidance;
+8. usefulness for later shot prompts.
+
+### Candidate Text Responsibility Rule
+
+For image generation, avoid relying on generated readable labels as the source of truth.
+
+Preferred split:
+
+- image model: visual board, views, silhouettes, clothing, props, empty label zones if useful;
+- project document: exact labels, review criteria, identity-lock fields;
+- later page/layout layer: any final readable Chinese labels if required.
+
+### Candidate Gate Sequence
+
+1. trial plan;
+2. execution brief;
+3. Skill-based prompt;
+4. main board generation;
+5. review and repair;
+6. transparent slices;
+7. identity lock;
+8. use in shot tests.
+
+## Trial Validation Questions
+
+This project should answer these questions before any rule becomes public knowledge.
+
+1. Can the three GPT character files be reliably transformed into a concise project execution brief?
+2. Does the `gpt-image-2` Advisor workflow produce a better prompt than direct drafting for this character board?
+3. Can one generated board maintain same-person consistency across front / side / back / walking views?
+4. Does the user's source photo help establish identity without turning the output into a portrait or passport-photo derivative?
+5. Are image-model text labels usable, or should all exact labels be handled outside the generated image?
+6. Does the accepted main board actually improve later five-shot test prompts?
+7. Which review criteria are general enough for future projects, and which are only specific to `在路上`?
+
+## Do Not Promote Yet
+
+The following should not be added to stable knowledge yet:
+
+- a final character-anchor directory standard;
+- a mandatory fixed board layout;
+- a universal prompt template;
+- a universal source-photo policy;
+- a claim that text labels should or should not be generated inside the image;
+- a claim that one board is enough before transparent slices.
+
+These can be considered only after this trial has produced and reviewed real images.
+
+## Immediate Next Step
+
+After this plan is approved, create:
+
+```text
+visual-system/characters/character-anchor-brief-v1.md
+```
+
+That brief should use the three GPT character intake files as primary sources, include the source-photo role, and prepare the material for a `gpt-image-2` Skill-based prompt document.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v1-codex-test.png b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v1-codex-test.png
new file mode 100644
index 0000000..a3ec337
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v1-codex-test.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-back-reference-crop-tight.png b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-back-reference-crop-tight.png
new file mode 100644
index 0000000..f89d65a
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-back-reference-crop-tight.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-codex-posture-test.png b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-codex-posture-test.png
new file mode 100644
index 0000000..1dfbd9f
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-codex-posture-test.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-front-reference-crop-tight.png b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-front-reference-crop-tight.png
new file mode 100644
index 0000000..37e6344
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-front-reference-crop-tight.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-front-reference-crop.png b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-front-reference-crop.png
new file mode 100644
index 0000000..b6725f8
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-front-reference-crop.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-side-reference-crop-tight.png b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-side-reference-crop-tight.png
new file mode 100644
index 0000000..c6d7fd0
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-side-reference-crop-tight.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-back-reference-crop-tight.png b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-back-reference-crop-tight.png
new file mode 100644
index 0000000..f4dfcf0
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-back-reference-crop-tight.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-side-reference-crop-tight.png b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-side-reference-crop-tight.png
new file mode 100644
index 0000000..39757f7
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-anchor-v2-walk-side-reference-crop-tight.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-chatgpt.png b/projects/2026-06-23-在路上/visual-system/characters/character-chatgpt.png
new file mode 100644
index 0000000..91bf016
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-chatgpt.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-identity-lock.md b/projects/2026-06-23-在路上/visual-system/characters/character-identity-lock.md
new file mode 100644
index 0000000..0c70bc0
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-identity-lock.md
@@ -0,0 +1,176 @@
+# Character Identity Lock
+
+## Status
+
+- Document status: user approved for transparent-slice generation.
+- Lock basis: user selected `character-anchor-v2-codex-posture-test.png` as the main character anchor board.
+- Execution status: identity lock only; do not generate new images directly from this file.
+- Next gated action: review generated transparent character slices before five-shot scene prompt work.
+
+## Accepted Anchor
+
+Accepted main character anchor board:
+
+```text
+visual-system/characters/character-anchor-v2-codex-posture-test.png
+```
+
+Acceptance reason:
+
+- V1 had the preferred character image direction.
+- V1 walking posture was too hunched.
+- V2 keeps the preferred traveler image while repairing the walk toward a straighter, chest-open, dignified posture.
+
+## Input Lineage
+
+Primary identity reference:
+
+```text
+C:\Users\wangq\Pictures\照片\photo.jpg
+```
+
+Role:
+
+- identity reference;
+- not an edit target;
+- not a passport-photo style target.
+
+Project visual references:
+
+- `visual-system/characters/character-anchor-v1-codex-test.png`: preferred character-image direction.
+- `visual-system/characters/character-chatgpt.png`: positive posture and board-quality reference.
+- `visual-system/characters/character-anchor-prompt-v2.md`: Skill-based prompt used for the accepted v2 board.
+- `visual-system/characters/character-anchor-review-v2.md`: review basis for accepting v2.
+
+## Core Identity
+
+The character is a 50-year-old male traveler, about 180cm tall.
+
+He is based on the user's identity reference, but translated into the `在路上` road-film figure:
+
+- weathered by life and systems;
+- warm but tired;
+- lonely but steady;
+- silent and not defeated;
+- tired but dignified;
+- capable of continuing forward.
+
+He is not a portrait-MV protagonist. In later shots, he should usually function as a moving coordinate inside a larger world.
+
+## Face, Hair, And Beard
+
+Locked traits:
+
+- black rectangular glasses;
+- dark, somewhat longer, slightly messy hair;
+- visible gray / white streaks are allowed and useful;
+- longer, denser beard / stubble, fuller than light stubble but still realistic and controlled;
+- weathered middle-aged face;
+- quiet, serious, tired expression with dignity.
+
+Avoid:
+
+- polished grooming;
+- celebrity portrait look;
+- glamour portrait expression;
+- passport-photo pose or lighting.
+
+## Body And Posture
+
+Locked traits:
+
+- about 180cm adult male body direction;
+- solid and capable, neither frail nor gym-styled;
+- normal standing posture: stable, weighted, restrained;
+- walking posture: upright chest, straighter spine, natural head and neck alignment, relaxed shoulders.
+
+Critical posture rule:
+
+> The traveler may look tired, but he must not walk with a bent waist, hunched back, rounded collapsed shoulders, bowed defeated head, or physically broken body.
+
+## Clothing And Props
+
+Locked clothing:
+
+- dark gray / dark khaki / coal-black long traveler's coat;
+- practical road-worn texture;
+- old but not ragged;
+- dark restrained inner layer;
+- old dark pants.
+
+Locked props:
+
+- old practical backpack;
+- medium volume;
+- worn / dusty but not broken;
+- stable enough to identify the character from behind.
+
+Locked footwear:
+
+- old walking boots;
+- durable, road-worn, able to carry mud;
+- not new fashion boots and not shoe-ad framing.
+
+## Wide-Shot Recognition
+
+Later MV shots may place the character at 1/8 to 1/12 of the frame.
+
+Recognition should survive primarily through:
+
+- upright walking silhouette;
+- long dark road-worn coat;
+- old backpack shape and placement;
+- boot / lower-body walking rhythm;
+- longer messy hair outline;
+- restrained, non-heroic gait.
+
+The face is useful in closer shots, but the later wide-shot system must not depend on facial detail.
+
+## Allowed Drift
+
+Allowed variations:
+
+- mud, dust, rain, wind, and road wear;
+- tired eyes;
+- slight pose variation;
+- restrained ruggedness;
+- small lighting or weather shifts;
+- different scene distances, including distant back and side views.
+
+## Forbidden Drift
+
+Do not drift into:
+
+- hunched walking posture;
+- bent waist;
+- rounded shoulders collapsing forward;
+- bowed defeated head;
+- collapsed, frail, or physically broken body;
+- homeless stereotype;
+- drunk stereotype;
+- pity image or spiritual defeat;
+- fashion model or travel blogger styling;
+- luxury outdoor gear;
+- heroic poster stance;
+- arms-open inspirational pose;
+- celebrity portrait;
+- glossy game concept art;
+- anime / cartoon styling;
+- fantasy, cyberpunk, neon, sci-fi, or futuristic clothing;
+- passport-photo style;
+- exact text dependence in generated boards.
+
+## Trial Notes
+
+This lock is still a trial-stage artifact.
+
+It verifies whether an accepted main character anchor board can be converted into a reusable identity contract before generating:
+
+- transparent character slices;
+- common visual style anchors;
+- five-shot test prompts for `s01`, `s10`, `s46`, `s47`, and `s51`.
+
+Remaining risk:
+
+- V2 proves the board-level identity and posture direction, but does not yet prove that wide environmental scene prompts can preserve the character identity.
+- That must be tested in the later five-shot run after this lock is reviewed.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v1-chroma-src.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v1-chroma-src.png
new file mode 100644
index 0000000..51b90f6
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v1-chroma-src.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v1.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v1.png
new file mode 100644
index 0000000..480dac7
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v1.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v2-from-anchor-preview.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v2-from-anchor-preview.png
new file mode 100644
index 0000000..946bbc3
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v2-from-anchor-preview.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v2-from-anchor.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v2-from-anchor.png
new file mode 100644
index 0000000..cb8fd7e
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-back-v2-from-anchor.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v1-chroma-src.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v1-chroma-src.png
new file mode 100644
index 0000000..8dca6aa
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v1-chroma-src.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v1.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v1.png
new file mode 100644
index 0000000..8ce493b
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v1.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v2-from-anchor-preview.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v2-from-anchor-preview.png
new file mode 100644
index 0000000..f4d6128
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v2-from-anchor-preview.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v2-from-anchor-raw-alpha.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v2-from-anchor-raw-alpha.png
new file mode 100644
index 0000000..d7468af
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v2-from-anchor-raw-alpha.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v2-from-anchor.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v2-from-anchor.png
new file mode 100644
index 0000000..3cc993e
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-front-v2-from-anchor.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v1-chroma-src.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v1-chroma-src.png
new file mode 100644
index 0000000..8aeb730
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v1-chroma-src.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v1.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v1.png
new file mode 100644
index 0000000..124128e
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v1.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v2-from-anchor-preview.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v2-from-anchor-preview.png
new file mode 100644
index 0000000..6ec0c2a
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v2-from-anchor-preview.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v2-from-anchor.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v2-from-anchor.png
new file mode 100644
index 0000000..5b19ada
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-side-v2-from-anchor.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v1-chroma-src.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v1-chroma-src.png
new file mode 100644
index 0000000..2f43e76
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v1-chroma-src.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v1.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v1.png
new file mode 100644
index 0000000..92d8d28
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v1.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor-preview.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor-preview.png
new file mode 100644
index 0000000..f1343cf
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor-preview.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png
new file mode 100644
index 0000000..32499ce
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-back-v2-from-anchor.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v1-chroma-src.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v1-chroma-src.png
new file mode 100644
index 0000000..627e13d
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v1-chroma-src.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v1.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v1.png
new file mode 100644
index 0000000..558d2af
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v1.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor-preview.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor-preview.png
new file mode 100644
index 0000000..7289349
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor-preview.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png
new file mode 100644
index 0000000..979f401
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-slice-walk-side-v2-from-anchor.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slice-front-repair-v2.md b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slice-front-repair-v2.md
new file mode 100644
index 0000000..ad27ad3
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slice-front-repair-v2.md
@@ -0,0 +1,105 @@
+# Character Transparent Slice Front Repair V2
+
+## Status
+
+- Review status: accepted as part of the v2-from-anchor transparent slice set.
+- Scope: front full-body slice only.
+- Repair reason: the first generated transparent slice set drifted away from the accepted v2 character anchor.
+- Repair strategy: derive the front slice directly from the accepted v2 anchor board instead of regenerating a new person from text.
+- Next gated action: completed by `character-transparent-slices-v2-from-anchor-review.md`.
+
+## Source
+
+Accepted anchor board:
+
+```text
+visual-system/characters/character-anchor-v2-codex-posture-test.png
+```
+
+Source crop:
+
+```text
+visual-system/characters/character-anchor-v2-front-reference-crop-tight.png
+```
+
+Final alpha PNG candidate:
+
+```text
+visual-system/characters/character-slice-front-v2-from-anchor.png
+```
+
+Preview:
+
+```text
+visual-system/characters/character-slice-front-v2-from-anchor-preview.png
+```
+
+## Execution Method
+
+This repair did not use independent prompt-only generation.
+
+Method:
+
+1. Crop the front full-body standing figure from the accepted v2 anchor board.
+2. Remove the light board background locally.
+3. Save the result as a transparent PNG.
+4. Preserve the old rejected v1 front slice as failed-trial evidence.
+
+Crop box from the v2 board:
+
+```text
+(20, 104, 264, 835)
+```
+
+Local processing environment:
+
+```text
+C:\Users\wangq\.cache\codex-runtimes\codex-primary-runtime\dependencies\python\python.exe
+```
+
+## Validation
+
+Final file:
+
+```text
+visual-system/characters/character-slice-front-v2-from-anchor.png
+```
+
+Validation result:
+
+| Check | Result |
+| --- | --- |
+| Mode | RGBA |
+| Size | 488x1440 |
+| Four corners transparent | Pass: `[0, 0, 0, 0]` |
+| Source identity | Directly derived from accepted v2 front figure |
+| New-person generation risk | Avoided |
+
+Alpha summary:
+
+```text
+transparent=247699
+partial=33986
+opaque=421035
+```
+
+## Review Notes
+
+Strengths:
+
+- Identity matches v2 because the figure is cropped from v2 itself.
+- Front face, glasses, hair, beard, coat, backpack straps, pants, and boots are preserved.
+- No new prompt-only person was synthesized.
+- Full body is retained.
+
+Known limitations:
+
+- Resolution is limited by the front cell inside the 16:9 anchor board.
+- Some soft edge cleanup is image-processing based, not model-native transparency.
+- This is a better identity anchor than the rejected v1 front slice, but not a high-resolution regenerated character render.
+
+## Recommendation
+
+Use this as the front transparent-slice candidate if identity fidelity is more important than maximum resolution.
+
+After user approval, repair `side` using the same principle: derive or reference-bind from the accepted v2 anchor board first, rather than text-only regeneration.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-preview-v1.png b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-preview-v1.png
new file mode 100644
index 0000000..4a85d50
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-preview-v1.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-prompt-v1.md b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-prompt-v1.md
new file mode 100644
index 0000000..1675553
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-prompt-v1.md
@@ -0,0 +1,325 @@
+# Character Transparent Slices Prompt V1
+
+## Status
+
+- Document status: used for first generation batch.
+- Execution status: five slice candidates generated; review pending.
+- Project: `在路上`.
+- Asset step: character transparent slices after accepted v2 main character anchor board.
+- Review file: `visual-system/characters/character-transparent-slices-review-v1.md`.
+
+## Purpose
+
+Create five transparent character cutout references for later MV shot prompting and possible compositing reference.
+
+The slices are not final MV frames. They are execution assets that should preserve the accepted character identity while removing scene background influence.
+
+Required slice set:
+
+| ID | Output candidate | Purpose |
+| --- | --- | --- |
+| `front` | front full body | Locks body proportion, coat front, face / hair / beard direction. |
+| `side` | side full body | Locks side silhouette, coat fall, backpack volume, profile. |
+| `back` | back full body | Locks back-view recognition without face. |
+| `walk-side` | walking side full body | Tests the upright walking posture required for road shots. |
+| `walk-back` | walking back full body | Tests the most important later wide-shot silhouette. |
+
+## Skill Workflow Record
+
+Mode check command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+Mode check 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)"
+}
+```
+
+Interpretation:
+
+- Use `gpt-image-2` as the prompt-engineering workflow.
+- Garden local generation is not enabled.
+- This Codex environment has a host image-generation path, so later execution should be treated as Mode B if the user approves generation.
+- No Garden scripts should be called for this step.
+
+Skill files actually 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\portraits-and-characters\character-sheet.md`
+- `C:\Users\wangq\.agents\skills\gpt-image-2\references\portraits-and-characters\pose-reference-sheet.md`
+
+Template adaptation:
+
+- `character-sheet.md` provides the front / side / back view logic.
+- `pose-reference-sheet.md` provides the same-character / pose-variation control logic.
+- This project adapts both into five independent transparent cutout prompts, not a full character board.
+
+## Source Files Actually Used
+
+Primary project sources:
+
+- `visual-system/characters/character-identity-lock.md`
+- `visual-system/characters/character-anchor-review-v2.md`
+- `visual-system/characters/character-anchor-prompt-v2.md`
+
+Visual references:
+
+- `visual-system/characters/character-anchor-v2-codex-posture-test.png`: accepted main character anchor board.
+- `visual-system/characters/character-anchor-v1-codex-test.png`: preferred original character-image direction.
+- `visual-system/characters/character-chatgpt.png`: positive upright walking posture / board-quality reference.
+
+Identity reference:
+
+```text
+C:\Users\wangq\Pictures\照片\photo.jpg
+```
+
+Role:
+
+- identity reference only;
+- not an edit target;
+- not a passport-photo style target;
+- use only to preserve the lived-person basis and face direction when the execution tool supports image references.
+
+## Asset Format
+
+Preferred output:
+
+- five separate PNG images;
+- transparent alpha background;
+- one full-body character per image;
+- no scene, no floor, no horizon, no labels, no board layout;
+- full body must be fully inside the canvas, including head, coat hem, backpack, hands, boots, and any stride extension;
+- consistent character scale across all five images.
+
+Preferred canvas:
+
+- individual vertical cutout canvas, about 2:3 or 3:4;
+- if the host tool only supports a horizontal frame, center the character and keep enough empty transparent margin for cropping.
+
+Transparency fallback:
+
+- If true alpha transparency is unavailable, use a completely flat, clean, uniform near-white or very light neutral matte background with no texture and no shadow; mark the result as non-transparent fallback in the review.
+
+## Character Lock For All Slices
+
+Every slice must depict the same person:
+
+- 50-year-old male traveler;
+- about 180cm tall;
+- black rectangular glasses;
+- dark, somewhat longer, slightly messy hair;
+- visible gray / white streaks allowed;
+- longer, denser beard / stubble, fuller than light stubble but still realistic and controlled;
+- weathered middle-aged face;
+- quiet, serious, tired expression with dignity;
+- solid adult body, capable of continuing forward;
+- dark gray / dark khaki / coal-black long traveler's coat;
+- dark restrained inner layer;
+- old dark pants;
+- old practical medium-volume backpack, worn / dusty but not broken;
+- old durable walking boots, road-worn and able to carry mud.
+
+Character temperament:
+
+- weathered by life and systems;
+- warm but tired;
+- lonely but steady;
+- silent and not defeated;
+- tired but dignified;
+- road-worn, not ragged.
+
+## Critical Posture Rule
+
+For all standing and walking slices:
+
+> The traveler may look tired, but he must not walk or stand with a bent waist, hunched back, rounded collapsed shoulders, bowed defeated head, or physically broken body.
+
+Walking slices must show:
+
+- upright chest;
+- straighter spine;
+- natural head and neck alignment;
+- relaxed shoulders that do not collapse forward;
+- steady forward movement;
+- grounded but dignified gait.
+
+## Shared Negative Constraints
+
+Strictly avoid:
+
+- hunched walking posture;
+- bent waist;
+- rounded shoulders collapsing forward;
+- bowed defeated head;
+- collapsed, frail, or physically broken body;
+- homeless stereotype;
+- drunk stereotype;
+- pity image or spiritual defeat;
+- fashion model styling;
+- polished travel blogger image;
+- luxury outdoor gear;
+- heroic poster stance;
+- arms-open inspirational pose;
+- celebrity portrait look;
+- glossy game concept art;
+- anime / cartoon styling;
+- fantasy armor or costume;
+- cyberpunk / neon / sci-fi / futuristic clothing;
+- passport-photo style;
+- front-facing glamour portrait;
+- scenery background;
+- city background;
+- muddy road background;
+- campfire background;
+- dramatic studio shadow;
+- text labels, captions, logos, UI, frames, borders, or annotation lines;
+- multiple different men instead of the same traveler.
+
+## Base Prompt Block
+
+Use this base block for all five slice prompts:
+
+```text
+Create one realistic full-body transparent PNG character cutout for the music-video project "在路上".
+
+Use the accepted character anchor board `character-anchor-v2-codex-posture-test.png` as the primary visual identity reference. Preserve the same 50-year-old male road traveler: about 180cm tall, black rectangular glasses, longer messy dark hair with possible gray/white streaks, longer and denser realistic beard/stubble, weathered middle-aged face, quiet tired dignity, solid adult body, dark gray / dark khaki / coal-black long road-worn traveler's coat, dark restrained inner layer, old dark pants, old practical medium-volume backpack, and old durable walking boots.
+
+Use the source photo only as an identity reference if image-reference input is supported. Do not copy its red background, passport-photo pose, suit, lighting, or portrait framing.
+
+The image is an isolated execution asset, not a final MV frame and not a character board. One full-body character only. Transparent alpha background. No scenery, no floor texture, no horizon, no text, no labels, no frame, no UI, no props other than the backpack and clothing system. Keep the whole body fully inside the canvas with clean transparent margins.
+
+Style: realistic cinematic pre-production character cutout, low-saturation documentary road-film feeling, practical and restrained, not anime, not cartoon, not glossy game concept art, not fashion editorial.
+
+The traveler may look tired, but he must not look collapsed, frail, homeless-coded, drunk-coded, heroic, or defeated.
+```
+
+## Slice Prompt 1: Front Full Body
+
+Output candidate:
+
+```text
+visual-system/characters/character-slice-front-v1.png
+```
+
+Prompt:
+
+```text
+[BASE PROMPT BLOCK]
+
+Pose requirement: front full-body standing view. The traveler faces the camera naturally, feet grounded, shoulders relaxed, coat front visible, backpack straps or side edges visible if natural, hands neutral and unobtrusive. The body should feel tall, solid, and capable of walking, not heroic and not posed for a fashion portrait.
+
+Purpose: lock age impression, body proportion, coat front, glasses, longer hair, fuller beard, and tired but dignified face direction.
+```
+
+## Slice Prompt 2: Side Full Body
+
+Output candidate:
+
+```text
+visual-system/characters/character-slice-side-v1.png
+```
+
+Prompt:
+
+```text
+[BASE PROMPT BLOCK]
+
+Pose requirement: clean side full-body standing view, profile direction, full head-to-boots silhouette visible. Show the coat fall, backpack thickness and placement, profile of glasses, hair outline, beard profile, and boot shape. The torso must not collapse forward. Keep the spine naturally upright and the head aligned.
+
+Purpose: lock side silhouette and backpack volume for later landscape-led side-view shots.
+```
+
+## Slice Prompt 3: Back Full Body
+
+Output candidate:
+
+```text
+visual-system/characters/character-slice-back-v1.png
+```
+
+Prompt:
+
+```text
+[BASE PROMPT BLOCK]
+
+Pose requirement: back full-body standing view. The traveler faces away from the camera. Show stable back-view recognition: longer messy hair outline, old medium backpack centered on the back, long dark coat from shoulders to lower body, old pants, and road-worn boots. Shoulders are relaxed but not folded forward.
+
+Purpose: lock the no-face identity used in distant MV shots.
+```
+
+## Slice Prompt 4: Walking Side Full Body
+
+Output candidate:
+
+```text
+visual-system/characters/character-slice-walk-side-v1.png
+```
+
+Prompt:
+
+```text
+[BASE PROMPT BLOCK]
+
+Pose requirement: side full-body walking view, one step forward, natural road-walking gait. The traveler walks with upright chest, straighter spine, natural head and neck alignment, and relaxed shoulders that do not collapse forward. The coat moves slightly with the step, backpack sits naturally, boots show grounded forward motion.
+
+Critical rule: do not bend the waist, do not hunch the back, do not bow the head, do not make him look physically broken or defeated. The walk should feel tired but dignified, steady, and capable of continuing.
+
+Purpose: lock the main side-walking silhouette for road shots.
+```
+
+## Slice Prompt 5: Walking Back Full Body
+
+Output candidate:
+
+```text
+visual-system/characters/character-slice-walk-back-v1.png
+```
+
+Prompt:
+
+```text
+[BASE PROMPT BLOCK]
+
+Pose requirement: back full-body walking view, the traveler moving away from the camera. Show a steady upright back, natural backpack placement, long coat movement, grounded boot stride, and the same longer hair outline. The body should read as continuing forward, not slumping away.
+
+Critical rule: shoulders must not fold forward, spine must not hunch, head must not droop in defeat. The back view should feel lonely and tired but still dignified.
+
+Purpose: lock the most important no-face walking silhouette for later wide shots.
+```
+
+## Review Checklist
+
+After generation, create a review file and check:
+
+1. Do all five slices depict the same traveler?
+2. Does the face / hair / beard / glasses system stay close to the accepted v2 anchor?
+3. Do coat, backpack, pants, and boots remain stable across all slices?
+4. Are the walking side and walking back slices upright enough?
+5. Are head, coat hem, hands, backpack, and boots fully inside the canvas?
+6. Is the background truly transparent, or only a fallback matte?
+7. Does the side view preserve backpack volume without torso collapse?
+8. Does the back view work without face details?
+9. Do the assets avoid fashion, hero, homeless, drunk, and defeated-body drift?
+10. Are the cutouts useful for later `s01`, `s10`, `s46`, `s47`, and `s51` prompt work?
+
+## Next Gated Action
+
+Only after user approval:
+
+1. generate the five slice candidates;
+2. save them under `visual-system/characters/`;
+3. create `character-transparent-slices-review-v1.md`;
+4. do not proceed to the common style anchor board until the slice review is accepted or intentionally skipped.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-review-v1.md b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-review-v1.md
new file mode 100644
index 0000000..7b61a55
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-review-v1.md
@@ -0,0 +1,152 @@
+# Character Transparent Slices Review V1
+
+## Status
+
+- Review status: not accepted by user.
+- Generation status: completed first five-slice batch.
+- Prompt/spec source: `visual-system/characters/character-transparent-slices-prompt-v1.md`.
+- Main identity source: accepted v2 character anchor board and `character-identity-lock.md`.
+- User rejection reason: the generated slices no longer look like the accepted v2 character anchor.
+- Next gated action: repair one slice at a time, starting from identity-critical front full body, before any style-anchor or five-shot work uses transparent character assets.
+
+## Generated Assets
+
+Final alpha PNG slices:
+
+| Slice | Final file | Source chroma file |
+| --- | --- | --- |
+| Front full body | `visual-system/characters/character-slice-front-v1.png` | `visual-system/characters/character-slice-front-v1-chroma-src.png` |
+| Side full body | `visual-system/characters/character-slice-side-v1.png` | `visual-system/characters/character-slice-side-v1-chroma-src.png` |
+| Back full body | `visual-system/characters/character-slice-back-v1.png` | `visual-system/characters/character-slice-back-v1-chroma-src.png` |
+| Walking side | `visual-system/characters/character-slice-walk-side-v1.png` | `visual-system/characters/character-slice-walk-side-v1-chroma-src.png` |
+| Walking back | `visual-system/characters/character-slice-walk-back-v1.png` | `visual-system/characters/character-slice-walk-back-v1-chroma-src.png` |
+
+Preview board:
+
+- `visual-system/characters/character-transparent-slices-preview-v1.png`
+
+## Execution Mode
+
+`gpt-image-2` Skill mode check:
+
+```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)"
+}
+```
+
+Execution interpretation:
+
+- Garden local generation was not enabled.
+- Host image generation was used as Mode B.
+- Each slice was generated separately.
+- The built-in image tool does not expose a direct output-path argument; generated defaults were copied into the project.
+- The built-in image tool also does not expose native transparent-background control, so each asset was generated on a flat green chroma-key background and converted locally to alpha PNG.
+
+## Generation Source Mapping
+
+Default host-generated source files were copied as follows:
+
+| Slice | Default source file |
+| --- | --- |
+| Front | `C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0cdcfc2a4d6685cc016a3a33d85a548191ace8cb7e4caa0623.png` |
+| Side | `C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0cdcfc2a4d6685cc016a3a341539308191833e4f10d980865e.png` |
+| Back | `C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0cdcfc2a4d6685cc016a3a34912710819196953af31c806bb2.png` |
+| Walk side | `C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0cdcfc2a4d6685cc016a3a34c91b3c819182e18394180e4000.png` |
+| Walk back | `C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_0cdcfc2a4d6685cc016a3a3503e1d481918e8786806eb9a450.png` |
+
+## Transparency Validation
+
+Local alpha conversion used:
+
+```text
+C:\Users\wangq\.codex\skills\.system\imagegen\scripts\remove_chroma_key.py
+```
+
+Python environment:
+
+```text
+C:\Users\wangq\.cache\codex-runtimes\codex-primary-runtime\dependencies\python\python.exe
+```
+
+Validation result:
+
+| File | Mode | Size | Transparent pixels | Partial pixels | Corner alpha |
+| --- | --- | --- | ---: | ---: | --- |
+| `character-slice-front-v1.png` | RGBA | 864x1821 | 959873 | 9220 | `[0, 0, 0, 0]` |
+| `character-slice-side-v1.png` | RGBA | 864x1821 | 1041950 | 8430 | `[0, 0, 0, 0]` |
+| `character-slice-back-v1.png` | RGBA | 864x1821 | 983848 | 8599 | `[0, 0, 0, 0]` |
+| `character-slice-walk-side-v1.png` | RGBA | 1024x1536 | 1195863 | 7495 | `[0, 0, 0, 0]` |
+| `character-slice-walk-back-v1.png` | RGBA | 864x1821 | 1044861 | 8060 | `[0, 0, 0, 0]` |
+
+Validation interpretation:
+
+- All five final files have an alpha channel.
+- All four corners are fully transparent.
+- Partial alpha pixels exist around edges, which is expected from soft matte extraction.
+
+## Visual Review
+
+Local preview file:
+
+```text
+visual-system/characters/character-transparent-slices-preview-v1.png
+```
+
+Observed strengths:
+
+- All five figures are full body and fully inside the canvas.
+- The long dark coat, backpack, old boots, longer hair, glasses, and fuller beard system remains coherent enough for this first slice batch.
+- The walking side pose is upright and does not repeat the rejected v1 hunched walking issue.
+- The walking back pose reads as continuing forward rather than collapsed.
+- Back-view identity is usable: backpack, long coat, hair outline, and boot stride are clear.
+
+Observed risks:
+
+- These were generated by host prompt-only Mode B. The local v2 anchor image and source photo were not attached as binary image inputs to the image tool.
+- Therefore, same-person fidelity is based on the approved textual identity lock plus generated-context continuity, not on a strict image-reference pipeline.
+- The front / side / back likeness is coherent at the character-system level, but should still be treated as a candidate set, not a final locked production asset set.
+- The side standing slice has a naturally forward profile and heavy coat; it should be checked by the user against the "not hunched" requirement.
+
+## Checklist
+
+| Check | Result | Notes |
+| --- | --- | --- |
+| Five requested slices exist | Pass | Front, side, back, walk-side, walk-back generated. |
+| Final files are alpha PNG | Pass | RGBA mode; transparent corners confirmed. |
+| Full body inside canvas | Pass | No obvious crop in local preview. |
+| Same character system | Pass with note | Coherent coat / backpack / boots / hair / beard system; not strict binary-reference identity. |
+| Walking side posture | Pass | Upright chest and straighter spine; not the rejected v1 hunch. |
+| Walking back posture | Pass | Continues forward without collapsed shoulders. |
+| Back-view recognition | Pass | Backpack, coat, hair outline, and boots are stable. |
+| Avoid fashion / hero drift | Pass | Practical road-worn silhouette, not fashion editorial. |
+| Transparency edge quality | Pass with note | Soft alpha edge present; user should inspect at final compositing scale. |
+
+## Recommendation
+
+Do not use this slice set for production prompts.
+
+Treat it as a failed identity-preservation test:
+
+1. Keep the files as failed-trial evidence.
+2. Do not delete them unless explicitly requested.
+3. Do not promote this transparent-slice workflow into reusable rules.
+4. Repair the slice workflow before generating more character execution assets.
+
+Root cause assessment:
+
+- The host tool produced five separate prompt-only generations.
+- The accepted v2 anchor and source photo were recorded as references, but were not attached as strict binary image inputs.
+- Therefore the model regenerated a similar "road traveler type" instead of preserving the accepted v2 person.
+
+Required process change:
+
+- Transparent character slices must be repaired one at a time.
+- The next attempt should be reference-bound or crop-derived from the accepted v2 board, not independent text-only generation.
+- The first repaired slice should be `front`, because face / hair / beard / glasses identity drift is easiest to judge there.
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-v2-from-anchor-preview.png b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-v2-from-anchor-preview.png
new file mode 100644
index 0000000..b2a0201
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-v2-from-anchor-preview.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-v2-from-anchor-review.md b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-v2-from-anchor-review.md
new file mode 100644
index 0000000..3a0ab48
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/characters/character-transparent-slices-v2-from-anchor-review.md
@@ -0,0 +1,98 @@
+# Character Transparent Slices V2 From Anchor Review
+
+## Status
+
+- Review status: accepted by user.
+- Scope: repaired transparent slice set derived from the accepted v2 character anchor board.
+- Repair reason: the first generated v1 slice set drifted away from the accepted v2 identity.
+- Repair strategy: crop each required pose from `character-anchor-v2-codex-posture-test.png` and remove the local board background.
+- Next gated action: proceed to the common style anchor board.
+
+## Source
+
+Accepted anchor board:
+
+```text
+visual-system/characters/character-anchor-v2-codex-posture-test.png
+```
+
+Preview board:
+
+```text
+visual-system/characters/character-transparent-slices-v2-from-anchor-preview.png
+```
+
+## Final Alpha PNG Assets
+
+| Slice | Final alpha PNG | Source crop |
+| --- | --- | --- |
+| Front | `visual-system/characters/character-slice-front-v2-from-anchor.png` | `visual-system/characters/character-anchor-v2-front-reference-crop-tight.png` |
+| Side | `visual-system/characters/character-slice-side-v2-from-anchor.png` | `visual-system/characters/character-anchor-v2-side-reference-crop-tight.png` |
+| Back | `visual-system/characters/character-slice-back-v2-from-anchor.png` | `visual-system/characters/character-anchor-v2-back-reference-crop-tight.png` |
+| Walk side | `visual-system/characters/character-slice-walk-side-v2-from-anchor.png` | `visual-system/characters/character-anchor-v2-walk-side-reference-crop-tight.png` |
+| Walk back | `visual-system/characters/character-slice-walk-back-v2-from-anchor.png` | `visual-system/characters/character-anchor-v2-walk-back-reference-crop-tight.png` |
+
+## Crop Boxes
+
+Crop boxes are from the accepted v2 board in pixel coordinates:
+
+| Slice | Crop box |
+| --- | --- |
+| Front | `(20, 104, 264, 835)` |
+| Side | `(276, 104, 474, 835)` |
+| Back | `(485, 104, 713, 835)` |
+| Walk side | `(948, 160, 1212, 835)` |
+| Walk back | `(1228, 120, 1412, 835)` |
+
+## Local Processing Method
+
+This repair did not use new image generation.
+
+Method:
+
+1. Crop each pose from the accepted v2 anchor board.
+2. Estimate the local neutral board background from crop borders.
+3. Remove border-connected light background.
+4. For walking crops, keep the largest connected foreground component to remove adjacent panel bleed.
+5. Apply a soft alpha matte.
+6. Save as alpha PNG.
+7. Build a checkerboard preview for human review.
+
+Processing environment:
+
+```text
+C:\Users\wangq\.cache\codex-runtimes\codex-primary-runtime\dependencies\python\python.exe
+```
+
+## Validation
+
+| Slice | Mode | Size | Transparent pixels | Partial pixels | Opaque pixels | Corner alpha |
+| --- | --- | --- | ---: | ---: | ---: | --- |
+| Front | RGBA | 488x1440 | 247699 | 33986 | 421035 | `[0, 0, 0, 0]` |
+| Side | RGBA | 396x1452 | 221077 | 26806 | 327109 | `[0, 0, 0, 0]` |
+| Back | RGBA | 456x1442 | 220243 | 32337 | 404972 | `[0, 0, 0, 0]` |
+| Walk side | RGBA | 470x1158 | 273011 | 31410 | 239839 | `[0, 0, 0, 0]` |
+| Walk back | RGBA | 368x1218 | 161412 | 20482 | 266330 | `[0, 0, 0, 0]` |
+
+## Review Notes
+
+Strengths:
+
+- Identity drift is avoided because every slice is derived directly from the accepted v2 board.
+- Front, side, back, walk-side, and walk-back all share the same v2 figure system.
+- The walk-side slice preserves the corrected upright walking posture from v2.
+- The walk-back slice no longer includes the stray neighboring-panel boot after connected-component cleanup.
+- All five final files have alpha channels and transparent corners.
+
+Known limitations:
+
+- Resolution is limited by the 16:9 anchor board cell size.
+- The side slice inherits the v2 board's panel crop and keeps the backpack close to the right edge.
+- Edge quality is local matte extraction quality, not native transparent rendering.
+- These are execution references for small MV figures, not high-resolution hero assets.
+
+## Recommendation
+
+Use this v2-from-anchor set if identity fidelity is the priority for small-figure MV shots.
+
+If user accepts this set, the next project step can move to the common style anchor board. If any slice needs repair, repair only that slice and keep the v2-derived method.
diff --git a/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-review-v1.md b/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-review-v1.md
new file mode 100644
index 0000000..d35bc55
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-review-v1.md
@@ -0,0 +1,68 @@
+# Style Anchor Board Review V1
+
+## Status
+
+- Document status: user accepted.
+- Project: `在路上`.
+- Board HTML: `visual-system/style/style-anchor-board-v1.html`.
+- Board PNG: `visual-system/style/style-anchor-board-v1.png`.
+- Render size: `1920x1080`.
+- Render method: deterministic HTML/CSS layout captured with local browser rendering.
+- Review result: accepted as the common style anchor board.
+
+## Source Images Used
+
+| Anchor | Image | Status |
+| --- | --- | --- |
+| A | `intake/imgs/A.风景主导型蓝图.png` | accepted source sample |
+| B | `intake/imgs/B.高位观察型蓝图.png` | accepted source sample |
+| C | `visual-system/style/style-anchor-c-repair-v1.png` | accepted repaired C sample |
+| D | `intake/imgs/D.夜路篝火型蓝图.png` | accepted source sample |
+
+Original C was not used as the board's C sample:
+
+```text
+intake/imgs/C.泥泞里的星空蓝图.png
+```
+
+## Local Inspection
+
+Pass:
+
+- A/B/C/D all render in the 2x2 sample grid.
+- C uses the accepted repaired muddy starlight image.
+- Chinese labels are rendered as real HTML text, not generated image text.
+- The right rules column is readable.
+- The bottom review standard is readable.
+- The board is 16:9 and exported as `1920x1080`.
+
+Watch points for user review:
+
+- A remains warmer because the accepted A source has sun at the horizon; this is already part of the accepted source image, not a board-rendering error.
+- The board is intentionally utilitarian and reference-oriented, not poster-like.
+
+## Acceptance Criteria Review
+
+| Criterion | Assessment | Notes |
+| --- | --- | --- |
+| A/B/C/D look like one MV | pass for review | Shared dark board, low-saturation sources, and consistent labels tie them together. |
+| First read is environment/world, not character poster | pass | The layout prioritizes four landscape/motif samples and rules. |
+| Chinese text exact and readable | pass | Text is deterministic HTML text. |
+| C uses repaired sample | pass | Uses `style-anchor-c-repair-v1.png`. |
+| Warmth remains restrained | mostly pass | D warmth is local; A source has horizon warmth but stays within the accepted A image. |
+| Rules column guides later prompts | pass | Color, light, composition, texture, and avoid rules are present. |
+| Usable for five-shot test setup | pass for review | It maps directly to S-01, S-10, S-46, S-47, and S-51 anchors. |
+
+## Decision
+
+User accepted:
+
+```text
+visual-system/style/style-anchor-board-v1.png
+```
+
+Next gated step:
+
+```text
+five-shot trial plan/spec
+```
diff --git a/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-spec-v1.md b/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-spec-v1.md
new file mode 100644
index 0000000..b7922c3
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-spec-v1.md
@@ -0,0 +1,320 @@
+# Style Anchor Board Spec V1
+
+## Status
+
+- Document status: user approved and rendered.
+- Execution status: rendered once as deterministic HTML/PNG; output accepted by user.
+- Project: `在路上`.
+- Purpose: define the common style anchor board layout using accepted A/B/C/D visual samples.
+- Output target after approval: deterministic layout artifact, not AI-generated text inside an image.
+
+Rendered outputs:
+
+```text
+visual-system/style/style-anchor-board-v1.html
+visual-system/style/style-anchor-board-v1.png
+```
+
+## Why This Is Not An Image-Generation Prompt
+
+The common style anchor board must contain exact Chinese labels and review rules.
+
+Therefore this board should be built as a deterministic layout artifact:
+
+- source images are inserted as image assets;
+- Chinese labels are rendered by HTML/CSS or another local layout method;
+- no generated model text is used as the source of truth;
+- the board can be revised without regenerating the visual samples.
+
+Recommended implementation after approval:
+
+```text
+visual-system/style/style-anchor-board-v1.html
+visual-system/style/style-anchor-board-v1.png
+```
+
+## Source Files Actually Used
+
+Primary sources:
+
+- `visual-system/style/style-anchor-brief-v1.md`
+- `visual-system/style/style-anchor-c-repair-review-v1.md`
+- `visual-system/visual-system.md`
+- `intake/阶段四-风格锚点板页面结构建议稿.md`
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+
+## Source Images
+
+All four source images are `1672x941`, approximately 16:9.
+
+| Anchor | Image | Role |
+| --- | --- | --- |
+| A | `intake/imgs/A.风景主导型蓝图.png` | Accepted A sample: landscape-led muddy road. |
+| B | `intake/imgs/B.高位观察型蓝图.png` | Accepted B sample: high-angle city-edge observation. |
+| C | `visual-system/style/style-anchor-c-repair-v1.png` | Accepted repaired C sample: muddy starlight. |
+| D | `intake/imgs/D.夜路篝火型蓝图.png` | Accepted D sample: restrained night-road fire. |
+
+Do not use the original C blueprint as the final C sample:
+
+```text
+intake/imgs/C.泥泞里的星空蓝图.png
+```
+
+It remains a problem reference only.
+
+## Board Format
+
+Canvas:
+
+- 16:9 horizontal.
+- Recommended render size: `1920x1080`.
+- Background: dark neutral coal gray, not pure black.
+- Overall mood: utilitarian film reference board, not a poster or marketing page.
+
+Background palette:
+
+- page background: coal gray / dark warm gray;
+- panel surfaces: slightly lighter dark gray;
+- rules text: off-white / muted gray;
+- accent: restrained warm orange only for small labels or D/fire references;
+- avoid bright blue/purple/cyber palette.
+
+## Layout Structure
+
+Use one MVP page.
+
+Page areas:
+
+1. Top definition band.
+2. Left main sample grid.
+3. Right compact rules column.
+4. Bottom review standard band.
+
+Recommended proportions:
+
+| Area | Approximate share | Notes |
+| --- | --- | --- |
+| Top band | 14% height | Title, style definition, composition rule. |
+| Main sample grid | 68% width, 72% height | A/B/C/D in a 2x2 grid. |
+| Right rules column | 26% width, 72% height | Color, light, composition, texture, avoid drift. |
+| Bottom band | 10% height | One-line review standard and short review questions. |
+
+## Top Definition Band
+
+Title:
+
+```text
+《在路上》MV 风格锚点板
+```
+
+Subtitle:
+
+```text
+Style Anchor Board｜Landscape Narrative Road Film
+```
+
+Style definition:
+
+```text
+现实纪实质感 + 克制的诗性超现实 + 公路纪录片式风景叙事
+```
+
+Composition rule:
+
+```text
+世界比人更大，路比人更长，人只是持续前行的坐标
+```
+
+## Sample Grid
+
+Use a 2x2 image grid on the left.
+
+Image behavior:
+
+- use each image as an inserted asset;
+- crop with `object-fit: cover`;
+- preserve each sample's main subject;
+- do not apply strong filters;
+- keep minor uniform dark overlay only if needed for labels;
+- labels must not cover the important subject area.
+
+### A Cell
+
+Title:
+
+```text
+A｜清晨泥泞旷野
+```
+
+Function:
+
+```text
+风景主导｜世界先于人出现
+```
+
+Avoid label:
+
+```text
+不要旅游片 / 英雄背影
+```
+
+### B Cell
+
+Title:
+
+```text
+B｜高位观察
+```
+
+Function:
+
+```text
+城市边缘｜系统里的小人物
+```
+
+Avoid label:
+
+```text
+不要城市宣传 / 科技感
+```
+
+### C Cell
+
+Title:
+
+```text
+C｜泥泞里的星空
+```
+
+Function:
+
+```text
+灵魂意象｜低处看见宇宙
+```
+
+Status tag:
+
+```text
+修正版已通过
+```
+
+Avoid label:
+
+```text
+不要日出 / 魔法 / 鞋靴广告
+```
+
+### D Cell
+
+Title:
+
+```text
+D｜夜路小火
+```
+
+Function:
+
+```text
+结尾温度｜远方也有人在走
+```
+
+Avoid label:
+
+```text
+不要露营 / 聚会 / 胜利
+```
+
+## Right Rules Column
+
+Keep this column compact and readable.
+
+### 色彩系统
+
+Text:
+
+```text
+冷灰 / 土褐 / 暗卡其 / 煤黑 / 湿地暗蓝灰
+暖色只作为情绪信号
+```
+
+### 光线系统
+
+Text:
+
+```text
+自然冷光、湿冷城市边缘、暗夜深度、小范围火光
+C 用泥水星点，不用日出日落
+```
+
+### 构图系统
+
+Text:
+
+```text
+环境优先，人物后置
+常规人物约 1/8-1/5，大远景可更小
+```
+
+### 质感系统
+
+Text:
+
+```text
+泥要湿，路要旧，城市要冷硬，火要小
+诗性来自现实质感，不来自炫光特效
+```
+
+### 禁止跑偏
+
+Text:
+
+```text
+人物写真 / 旅行广告 / 城市宣传 / 赛博科技 / 露营聚会 / 胜利海报
+```
+
+## Bottom Review Band
+
+One-line review standard:
+
+```text
+第一眼看到的应该是路、世界、泥、夜色和火光；人物只是其中继续前行的坐标。
+```
+
+Short review questions:
+
+```text
+像同一支 MV 吗？人物是否过大？诗性是否来自现实质感？是否仍然在继续走？
+```
+
+## Acceptance Criteria
+
+The rendered board passes only if:
+
+1. A/B/C/D look like one MV, not four unrelated projects.
+2. The first read is environment and road-film world, not a character poster.
+3. Chinese text is exact, readable, and not model-generated.
+4. C uses `style-anchor-c-repair-v1.png`, not the original warm C blueprint.
+5. Warmth remains local and restrained, mainly belonging to D.
+6. The rules column is compact enough to guide later shot prompts.
+7. The board is usable as a visual dictionary for S-01, S-10, S-46, S-47, and S-51.
+
+## Next Gated Action
+
+This spec has been approved and rendered as:
+
+```text
+visual-system/style/style-anchor-board-v1.html
+visual-system/style/style-anchor-board-v1.png
+```
+
+Review file:
+
+```text
+visual-system/style/style-anchor-board-review-v1.md
+```
+
+The rendered style anchor board has been reviewed and accepted. Next work should start with a five-shot trial plan/spec before any shot image generation.
diff --git a/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.html b/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.html
new file mode 100644
index 0000000..58c41fd
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.html
@@ -0,0 +1,385 @@
+<!doctype html>
+<html lang="zh-CN">
+<head>
+  <meta charset="utf-8">
+  <meta name="viewport" content="width=device-width, initial-scale=1">
+  <title>《在路上》MV 风格锚点板</title>
+  <style>
+    :root {
+      --bg: #171716;
+      --panel: #222220;
+      --panel-2: #2b2a27;
+      --line: #3a3934;
+      --text: #ece8dd;
+      --muted: #b6b0a4;
+      --dim: #8d877d;
+      --accent: #c88745;
+      --cold: #8fa2aa;
+      --danger: #d7b179;
+    }
+
+    * {
+      box-sizing: border-box;
+    }
+
+    html,
+    body {
+      margin: 0;
+      width: 100%;
+      min-height: 100%;
+      background: var(--bg);
+      font-family: "Microsoft YaHei", "Noto Sans CJK SC", "PingFang SC", "Segoe UI", sans-serif;
+      color: var(--text);
+    }
+
+    body {
+      display: grid;
+      place-items: start center;
+    }
+
+    .board {
+      width: 1920px;
+      height: 1080px;
+      padding: 28px;
+      display: grid;
+      grid-template-rows: 132px 770px 88px;
+      gap: 18px;
+      overflow: hidden;
+      background:
+        linear-gradient(135deg, rgba(255, 255, 255, 0.035), rgba(255, 255, 255, 0) 36%),
+        radial-gradient(circle at 80% 14%, rgba(199, 135, 69, 0.10), transparent 25%),
+        var(--bg);
+    }
+
+    .top {
+      border: 1px solid var(--line);
+      background: rgba(34, 34, 32, 0.86);
+      display: grid;
+      grid-template-columns: 1fr auto;
+      align-items: center;
+      padding: 22px 28px;
+    }
+
+    .title {
+      display: flex;
+      flex-direction: column;
+      gap: 10px;
+      min-width: 0;
+    }
+
+    h1 {
+      margin: 0;
+      font-size: 44px;
+      line-height: 1;
+      font-weight: 700;
+      letter-spacing: 0;
+    }
+
+    .subtitle {
+      margin: 0;
+      color: var(--muted);
+      font-size: 20px;
+      line-height: 1.2;
+      letter-spacing: 0;
+    }
+
+    .principles {
+      display: grid;
+      gap: 10px;
+      text-align: right;
+      max-width: 890px;
+      min-width: 760px;
+      font-size: 23px;
+      line-height: 1.22;
+      font-weight: 600;
+      color: var(--text);
+    }
+
+    .principles span:last-child {
+      color: var(--cold);
+    }
+
+    .main {
+      display: grid;
+      grid-template-columns: 1340px 502px;
+      gap: 22px;
+      min-height: 0;
+    }
+
+    .samples {
+      display: grid;
+      grid-template-columns: repeat(2, 1fr);
+      grid-template-rows: repeat(2, 1fr);
+      gap: 18px;
+      min-height: 0;
+    }
+
+    .sample {
+      border: 1px solid var(--line);
+      background: var(--panel);
+      display: grid;
+      grid-template-rows: 272px 1fr;
+      overflow: hidden;
+      min-width: 0;
+      min-height: 0;
+    }
+
+    .image-wrap {
+      position: relative;
+      min-height: 0;
+      overflow: hidden;
+      background: #111;
+    }
+
+    .image-wrap img {
+      width: 100%;
+      height: 100%;
+      object-fit: cover;
+      display: block;
+    }
+
+    .anchor-mark {
+      position: absolute;
+      top: 14px;
+      left: 14px;
+      padding: 6px 10px;
+      border: 1px solid rgba(236, 232, 221, 0.24);
+      background: rgba(17, 17, 16, 0.70);
+      color: var(--text);
+      font-size: 18px;
+      line-height: 1;
+      font-weight: 700;
+    }
+
+    .status {
+      position: absolute;
+      top: 14px;
+      right: 14px;
+      padding: 6px 10px;
+      border: 1px solid rgba(199, 135, 69, 0.48);
+      background: rgba(37, 28, 22, 0.76);
+      color: var(--danger);
+      font-size: 16px;
+      line-height: 1;
+      font-weight: 700;
+    }
+
+    .sample-text {
+      padding: 12px 16px 11px;
+      display: grid;
+      grid-template-columns: 1fr;
+      align-content: center;
+      gap: 4px;
+      min-width: 0;
+    }
+
+    .sample h2 {
+      margin: 0;
+      font-size: 24px;
+      line-height: 1.1;
+      letter-spacing: 0;
+    }
+
+    .function {
+      margin: 0;
+      color: var(--muted);
+      font-size: 16px;
+      line-height: 1.2;
+      white-space: nowrap;
+      overflow: hidden;
+      text-overflow: ellipsis;
+    }
+
+    .avoid {
+      margin: 0;
+      color: var(--accent);
+      font-size: 15px;
+      line-height: 1.2;
+      white-space: nowrap;
+      overflow: hidden;
+      text-overflow: ellipsis;
+    }
+
+    .rules {
+      border: 1px solid var(--line);
+      background: rgba(34, 34, 32, 0.90);
+      padding: 18px;
+      display: grid;
+      grid-template-rows: repeat(5, 1fr);
+      gap: 13px;
+      min-height: 0;
+    }
+
+    .rule {
+      border: 1px solid rgba(58, 57, 52, 0.92);
+      background: rgba(43, 42, 39, 0.86);
+      padding: 16px 17px;
+      display: grid;
+      align-content: center;
+      gap: 9px;
+      min-height: 0;
+    }
+
+    .rule h3 {
+      margin: 0;
+      font-size: 22px;
+      line-height: 1.1;
+      letter-spacing: 0;
+      color: var(--text);
+    }
+
+    .rule p {
+      margin: 0;
+      color: var(--muted);
+      font-size: 17px;
+      line-height: 1.45;
+      letter-spacing: 0;
+    }
+
+    .rule strong {
+      color: var(--accent);
+      font-weight: 700;
+    }
+
+    .bottom {
+      border: 1px solid var(--line);
+      background: rgba(34, 34, 32, 0.86);
+      display: grid;
+      grid-template-columns: 1.18fr 1fr;
+      align-items: center;
+      gap: 28px;
+      padding: 16px 24px;
+    }
+
+    .review-standard {
+      margin: 0;
+      font-size: 25px;
+      line-height: 1.25;
+      font-weight: 700;
+      color: var(--text);
+    }
+
+    .review-standard span {
+      color: var(--accent);
+    }
+
+    .questions {
+      display: grid;
+      grid-template-columns: repeat(2, minmax(0, 1fr));
+      gap: 8px 14px;
+      margin: 0;
+      padding: 0;
+      list-style: none;
+      color: var(--muted);
+      font-size: 17px;
+      line-height: 1.25;
+    }
+
+    .questions li::before {
+      content: "· ";
+      color: var(--accent);
+      font-weight: 700;
+    }
+  </style>
+</head>
+<body>
+  <main class="board" aria-label="《在路上》MV 风格锚点板">
+    <header class="top">
+      <div class="title">
+        <h1>《在路上》MV 风格锚点板</h1>
+        <p class="subtitle">Style Anchor Board｜Landscape Narrative Road Film</p>
+      </div>
+      <div class="principles">
+        <span>现实纪实质感 + 克制的诗性超现实 + 公路纪录片式风景叙事</span>
+        <span>世界比人更大，路比人更长，人只是持续前行的坐标</span>
+      </div>
+    </header>
+
+    <section class="main">
+      <div class="samples">
+        <article class="sample">
+          <div class="image-wrap">
+            <img src="../../intake/imgs/A.风景主导型蓝图.png" alt="A 清晨泥泞旷野">
+            <div class="anchor-mark">A</div>
+          </div>
+          <div class="sample-text">
+            <h2>A｜清晨泥泞旷野</h2>
+            <p class="function">风景主导｜世界先于人出现</p>
+            <p class="avoid">不要旅游片 / 英雄背影</p>
+          </div>
+        </article>
+
+        <article class="sample">
+          <div class="image-wrap">
+            <img src="../../intake/imgs/B.高位观察型蓝图.png" alt="B 高位观察">
+            <div class="anchor-mark">B</div>
+          </div>
+          <div class="sample-text">
+            <h2>B｜高位观察</h2>
+            <p class="function">城市边缘｜系统里的小人物</p>
+            <p class="avoid">不要城市宣传 / 科技感</p>
+          </div>
+        </article>
+
+        <article class="sample">
+          <div class="image-wrap">
+            <img src="style-anchor-c-repair-v1.png" alt="C 泥泞里的星空">
+            <div class="anchor-mark">C</div>
+            <div class="status">修正版已通过</div>
+          </div>
+          <div class="sample-text">
+            <h2>C｜泥泞里的星空</h2>
+            <p class="function">灵魂意象｜低处看见宇宙</p>
+            <p class="avoid">不要日出 / 魔法 / 鞋靴广告</p>
+          </div>
+        </article>
+
+        <article class="sample">
+          <div class="image-wrap">
+            <img src="../../intake/imgs/D.夜路篝火型蓝图.png" alt="D 夜路小火">
+            <div class="anchor-mark">D</div>
+          </div>
+          <div class="sample-text">
+            <h2>D｜夜路小火</h2>
+            <p class="function">结尾温度｜远方也有人在走</p>
+            <p class="avoid">不要露营 / 聚会 / 胜利</p>
+          </div>
+        </article>
+      </div>
+
+      <aside class="rules" aria-label="风格规则区">
+        <section class="rule">
+          <h3>色彩系统</h3>
+          <p>冷灰 / 土褐 / 暗卡其 / 煤黑 / 湿地暗蓝灰<br><strong>暖色只作为情绪信号</strong></p>
+        </section>
+        <section class="rule">
+          <h3>光线系统</h3>
+          <p>自然冷光、湿冷城市边缘、暗夜深度、小范围火光<br>C 用泥水星点，不用日出日落</p>
+        </section>
+        <section class="rule">
+          <h3>构图系统</h3>
+          <p>环境优先，人物后置<br>常规人物约 1/8-1/5，大远景可更小</p>
+        </section>
+        <section class="rule">
+          <h3>质感系统</h3>
+          <p>泥要湿，路要旧，城市要冷硬，火要小<br>诗性来自现实质感，不来自炫光特效</p>
+        </section>
+        <section class="rule">
+          <h3>禁止跑偏</h3>
+          <p>人物写真 / 旅行广告 / 城市宣传 / 赛博科技 / 露营聚会 / 胜利海报</p>
+        </section>
+      </aside>
+    </section>
+
+    <footer class="bottom">
+      <p class="review-standard">第一眼看到的应该是<span>路、世界、泥、夜色和火光</span>；人物只是其中继续前行的坐标。</p>
+      <ul class="questions">
+        <li>像同一支 MV 吗？</li>
+        <li>人物是否过大？</li>
+        <li>诗性是否来自现实质感？</li>
+        <li>是否仍然在继续走？</li>
+      </ul>
+    </footer>
+  </main>
+</body>
+</html>
diff --git a/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png b/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png
new file mode 100644
index 0000000..59bea3a
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/style/style-anchor-board-v1.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/style/style-anchor-brief-v1.md b/projects/2026-06-23-在路上/visual-system/style/style-anchor-brief-v1.md
new file mode 100644
index 0000000..80b835a
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/style/style-anchor-brief-v1.md
@@ -0,0 +1,466 @@
+# Style Anchor Brief V1
+
+## Status
+
+- Document status: user approved; C repair candidate accepted.
+- Execution status: brief used as source for C repair prompt/spec and accepted C sample.
+- Project: `在路上`.
+- Source basis: accepted visual system, style anchor trial plan, and A/B/C/D blueprint images.
+- Next gated action: review `style-anchor-board-spec-v1.md` before rendering the common style anchor board.
+
+## Purpose
+
+This brief defines the project-specific common style anchor requirements for `在路上`.
+
+It is not a final image prompt.
+
+It should answer:
+
+> What kind of image belongs to this MV, and what kind of image is good-looking but wrong for this MV?
+
+The style anchor must keep later wilderness, city-edge, muddy-puddle, night-road, and firelight shots inside one coherent visual world.
+
+## Source Files Actually Used
+
+Primary style sources:
+
+- `visual-system/style/style-anchor-trial-plan.md`
+- `visual-system/visual-system.md`
+- `intake/阶段四-通用表达与视觉系统设计.md`
+- `intake/阶段四-风格锚点板页面结构建议稿.md`
+
+Execution control:
+
+- `project.md`
+- `execution-plan.md`
+
+Source images:
+
+- `intake/imgs/A.风景主导型蓝图.png`
+- `intake/imgs/B.高位观察型蓝图.png`
+- `intake/imgs/C.泥泞里的星空蓝图.png`
+- `intake/imgs/D.夜路篝火型蓝图.png`
+
+Local inspection aid:
+
+- `visual-system/style/style-blueprints-contactsheet.png`
+
+## Current Source Image Roles
+
+| Anchor | Source image | Status | Use |
+| --- | --- | --- | --- |
+| A | `intake/imgs/A.风景主导型蓝图.png` | usable base | Main landscape-led road composition. |
+| B | `intake/imgs/B.高位观察型蓝图.png` | usable base | High-angle city-edge observation. |
+| C | `visual-system/style/style-anchor-c-repair-v1.png` | accepted repair sample | Final C visual sample for the common style anchor board. |
+| C original | `intake/imgs/C.泥泞里的星空蓝图.png` | problem reference | Retain only as earlier composition/problem reference. |
+| D | `intake/imgs/D.夜路篝火型蓝图.png` | usable base | Restrained night-road fire warmth. |
+
+Important distinction:
+
+- A/B/D can be used as style samples in the eventual board.
+- C has been rebuilt and accepted as `visual-system/style/style-anchor-c-repair-v1.png`.
+
+## Core Style Definition
+
+The MV visual style is:
+
+> Real documentary texture + restrained poetic surrealism + landscape-led road-film composition.
+
+Chinese working definition:
+
+> 现实纪实质感 + 克制的诗性超现实 + 公路纪录片式风景叙事。
+
+Core composition rule:
+
+> The world is larger than the person, the road is longer than the person, and the traveler is only the coordinate that keeps moving.
+
+Chinese working rule:
+
+> 世界比人更大，路比人更长，人只是持续前行的坐标。
+
+The emotional center is not victory. It is continuation.
+
+## Style Anchor A: Landscape-Led Road
+
+Working title:
+
+> A｜清晨泥泞旷野：世界先于人出现
+
+Function:
+
+- main composition standard;
+- proves world-before-person;
+- establishes the road-film grammar for the whole MV.
+
+Current source:
+
+```text
+intake/imgs/A.风景主导型蓝图.png
+```
+
+Preserve:
+
+- long road;
+- large world;
+- small person;
+- mud / frost / rough terrain;
+- cold morning or pre-dawn feeling;
+- low saturation;
+- loneliness without collapse.
+
+Avoid:
+
+- travel advertisement;
+- scenic tourism image;
+- hero backshot;
+- large portrait subject;
+- polished inspirational poster.
+
+Use in later tests:
+
+- `S-01` opening landscape;
+- any shot where the first impression must be road / world / continuation.
+
+## Style Anchor B: High-Angle City Edge
+
+Working title:
+
+> B｜高位观察：系统边缘里的小人物
+
+Function:
+
+- city-edge and system-pressure style;
+- observation rather than portrait;
+- proves that urban shots still belong to the same road-film world.
+
+Current source:
+
+```text
+intake/imgs/B.高位观察型蓝图.png
+```
+
+Preserve:
+
+- high-angle observation;
+- city edge, not city center;
+- wet ground;
+- rough terrain;
+- steel / concrete / distant skyline pressure;
+- small traveler;
+- documentary distance.
+
+Avoid:
+
+- city promotion image;
+- clean commercial skyline;
+- futuristic UI / code / neon / cyber city;
+- tech-product aesthetic;
+- ordinary tourism aerial shot.
+
+Use in later tests:
+
+- `S-10` steel canyon / system-pressure shot.
+
+## Style Anchor C: Muddy Starlight
+
+Working title:
+
+> C｜泥泞里的星空：低处看见宇宙
+
+Function:
+
+- soul image of the MV;
+- must prove that poetry comes from real mud, not from fantasy effects;
+- must connect low place and high place through broken starlight reflection.
+
+Original source:
+
+```text
+intake/imgs/C.泥泞里的星空蓝图.png
+```
+
+Current status:
+
+- repaired and accepted as final C sample for the common style anchor board;
+- original C source remains only as a problem reference and composition hint.
+
+Accepted repair sample:
+
+```text
+visual-system/style/style-anchor-c-repair-v1.png
+```
+
+What can be retained from the current C blueprint:
+
+- low mud-level perspective;
+- real muddy texture;
+- puddle as main visual subject;
+- boot / body element only at the edge.
+
+Problems to fix:
+
+- too close to sunrise / sunset reflection;
+- too warm and bright;
+- sky / star reflection is not broken and visible enough;
+- the spiritual function is weaker than required.
+
+Required rebuilt C direction:
+
+- dark night or pre-dawn;
+- muddy puddle is the visual subject;
+- visible broken star reflection in the water;
+- real mud around the puddle;
+- low-place-sees-high-place feeling;
+- old boot or traveler element only at the edge, under about 15% of the image;
+- restrained, quiet, documentary-real, not fantasy.
+
+Strictly avoid:
+
+- sunrise;
+- sunset;
+- broad warm horizon glow;
+- pretty reflection only;
+- fantasy magic;
+- glowing portal / cosmic effects;
+- shoe advertisement framing;
+- only mud and no stars;
+- only stars and no mud.
+
+Use in later tests:
+
+- `S-46` muddy starlight.
+
+Gate:
+
+- C repair prompt/spec has been completed and accepted before final common style board assembly.
+
+## Style Anchor D: Night Road Fire
+
+Working title:
+
+> D｜夜路小火：远方也有人在走
+
+Function:
+
+- ending warmth anchor;
+- confirms the presence of other people without turning into reunion or victory;
+- defines restrained firelight logic for late shots.
+
+Current source:
+
+```text
+intake/imgs/D.夜路篝火型蓝图.png
+```
+
+Preserve:
+
+- dark road;
+- small fire;
+- local warm-orange light;
+- distant faint fire points or lights;
+- restrained night depth;
+- loneliness with connection.
+
+Avoid:
+
+- camping advertisement;
+- campfire party;
+- group reunion;
+- victory image;
+- outdoor lifestyle promotion;
+- large theatrical firelight.
+
+Use in later tests:
+
+- `S-47` restrained near-fire transition;
+- `S-51` final wide with distant fire.
+
+## Color System
+
+Base palette:
+
+- cold gray;
+- earth brown;
+- dark khaki;
+- coal black;
+- wet dark blue-gray.
+
+Emotional colors:
+
+- small warm-orange fire;
+- weak pre-dawn silver-blue;
+- very small pale-gold edge only if needed.
+
+Rules:
+
+- low saturation most of the time;
+- warmth should be a signal, not the general palette;
+- dark scenes must retain depth;
+- C must stay away from broad warm sunrise / sunset light.
+
+Avoid:
+
+- bright tourism palette;
+- high-saturation advertising palette;
+- cyber blue / purple;
+- sweet warm yellow;
+- heavy commercial orange-blue blockbuster grading.
+
+## Light System
+
+Preferred light:
+
+- overcast;
+- cold morning;
+- thin frost / pre-dawn;
+- wet city gray;
+- bridge shadow;
+- dusk;
+- dark night;
+- small local firelight.
+
+Rules:
+
+- daylight should feel natural and restrained;
+- city edge should feel wet, gray, and cold;
+- night should remain dark;
+- fire should be local and small;
+- muddy starlight should use low reflection and broken star points, not sun.
+
+Avoid:
+
+- bright noon;
+- tourist sunlight;
+- broad theatrical rim light;
+- stage-like night lighting;
+- beautiful sunset as shortcut for emotion.
+
+## Composition System
+
+Global rule:
+
+> Environment first, traveler second.
+
+Traveler scale:
+
+- ordinary shots: about 1/8 to 1/5 of the frame;
+- final wide or large landscape: about 1/20 to 1/12;
+- C muddy starlight: boot / body element under about 15%, if present.
+
+Preferred compositions:
+
+- long horizontal road;
+- traveler lower or side-positioned;
+- large sky, ground, road, mud, or city-edge space;
+- high-angle observation;
+- low mud-level view;
+- distant fire points;
+- wide shot and medium-wide shot.
+
+Avoid:
+
+- centered hero portrait;
+- front-face half-body portrait as a default;
+- low-angle heroic pose;
+- travel blogger backshot;
+- campfire gathering center composition;
+- shoe closeup as product image.
+
+## Texture System
+
+Real texture anchors:
+
+- mud is wet and heavy;
+- puddles are irregular and dirty;
+- road is long and worn;
+- city edge is cold and hard;
+- fire is small and real;
+- boots and backpack are old and used.
+
+Poetry should come from:
+
+- low-place reflection;
+- wind and empty space;
+- small fire in night;
+- broken starlight in muddy water;
+- the traveler being small inside a larger world.
+
+Avoid:
+
+- over-clean surfaces;
+- over-polished cinematic beauty;
+- fantasy effects;
+- commercial product texture;
+- pure dirt / despair without poetic opening.
+
+## Final Style Anchor Board Requirements
+
+The final common style anchor board should be a layout artifact, not a single generated image with embedded text.
+
+Reason:
+
+- exact Chinese labels must be reliable;
+- generated text should not carry the source of truth;
+- A/B/D already exist as visual sources;
+- C must be repaired first.
+
+Recommended board:
+
+- one 16:9 horizontal page;
+- dark gray / coal-gray background, not pure black;
+- title and two top rules;
+- four visual sample cells A/B/C/D;
+- C marked as repaired or priority-fix depending on stage;
+- right-side compact rules: color, light, composition, texture, avoid drift;
+- bottom review rule line.
+
+Minimum content:
+
+1. `《在路上》MV 风格锚点板`;
+2. `现实纪实质感 + 克制的诗性超现实 + 公路纪录片式风景叙事`;
+3. `世界比人更大，路比人更长，人只是持续前行的坐标`;
+4. A/B/C/D samples;
+5. color / light / composition / texture rules;
+6. C marked as repaired / accepted;
+7. one-line review standard.
+
+## One-Line Review Standard
+
+Use this as the bottom rule of the final board:
+
+> 第一眼看到的应该是路、世界、泥、夜色和火光；人物只是其中继续前行的坐标。
+
+## Review Questions
+
+Use these questions after C repair and after final board assembly.
+
+1. Do A/B/C/D look like the same MV, not four separate projects?
+2. Does A establish road and world before person?
+3. Does B feel like documentary observation at the system edge?
+4. Does C truly show muddy starlight, not sunrise or pretty reflection?
+5. Does D feel warm but restrained, not camping or reunion?
+6. Does the palette stay low-saturation, with warmth only as a signal?
+7. Does the board guide later shot prompts, not just look attractive?
+
+## Next Gated Action
+
+The brief has already been used to create:
+
+```text
+visual-system/style/style-anchor-c-repair-prompt-v1.md
+```
+
+The prompt/spec has produced one accepted C repair sample:
+
+```text
+visual-system/style/style-anchor-c-repair-v1.png
+visual-system/style/style-anchor-c-repair-review-v1.md
+```
+
+Next decision:
+
+- review the common style anchor board specification before rendering the final board.
+
+```text
+visual-system/style/style-anchor-board-spec-v1.md
+```
diff --git a/projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-prompt-v1.md b/projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-prompt-v1.md
new file mode 100644
index 0000000..0225846
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-prompt-v1.md
@@ -0,0 +1,237 @@
+# Style Anchor C Repair Prompt V1
+
+## Status
+
+- Document status: user approved for one C repair generation.
+- Execution status: used once in Mode B host generation; output accepted by user via `style-anchor-c-repair-review-v1.md`.
+- Project: `在路上`.
+- Anchor: C｜泥泞里的星空.
+- Purpose: rebuild the C style anchor as a dark muddy starlight image.
+- Next gated action: completed; accepted C repair was used in the common style anchor board.
+
+Generated candidate:
+
+```text
+visual-system/style/style-anchor-c-repair-v1.png
+```
+
+## Skill Mode Check
+
+Command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+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)"
+}
+```
+
+Interpretation:
+
+- Use `gpt-image-2` as the prompt-engineering workflow.
+- Garden local generation is not enabled.
+- This Codex environment has a host image-generation path, so later execution should be treated as Mode B if the user approves generation.
+- Do not call Garden scripts for this step.
+
+## Skill Files Actually 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`
+
+Template adaptation:
+
+- `concept-scene.md` is used for foreground / midground / background / lighting structure.
+- `minimalist-mood-scene.md` is used for restraint, negative space, and mood-first control.
+- The final target is not epic concept art and not illustration. It should read as a realistic cinematic documentary still.
+
+## Source Files Actually Used
+
+Primary project sources:
+
+- `visual-system/style/style-anchor-brief-v1.md`
+- `visual-system/style/style-anchor-trial-plan.md`
+- `visual-system/visual-system.md`
+- `intake/阶段四-通用表达与视觉系统设计.md`
+- `intake/阶段四-风格锚点板页面结构建议稿.md`
+
+Problem reference image:
+
+```text
+intake/imgs/C.泥泞里的星空蓝图.png
+```
+
+Role:
+
+- problem reference;
+- composition hint;
+- not accepted target;
+- not an edit target unless the execution tool explicitly supports image editing and the user asks for edit mode.
+
+What to retain from the problem reference:
+
+- mud-level camera;
+- real muddy road texture;
+- puddle as main visual subject;
+- old boot or traveler trace only at the edge;
+- low place looking upward through reflection.
+
+What to reject from the problem reference:
+
+- sunrise / sunset feeling;
+- broad warm horizon glow;
+- pretty orange reflection;
+- overly bright sky reflection;
+- weak or secondary star reflection.
+
+## C Repair Target
+
+The repaired C image must answer:
+
+> 人还在泥里，但他终于在泥里看见了星空。
+
+The image should not feel like escape from mud. It should feel like meaning discovered inside the low place.
+
+Required direction:
+
+- dark night or deep pre-dawn;
+- real muddy puddle as the main subject;
+- visible broken star reflection in the water;
+- real wet mud around the puddle;
+- old boot or traveler element only at the edge, under about 15% of the frame;
+- quiet and restrained;
+- documentary-real with poetic pressure;
+- 16:9 horizontal.
+
+## Rendered Prompt
+
+Use the existing C blueprint only as a problem reference and composition hint. Preserve its useful parts: ground-level mud perspective, irregular muddy puddle as the main subject, real wet mud texture, and a small old boot or traveler trace at the edge. Do not preserve its sunrise / sunset warmth, broad orange horizon glow, or pretty warm reflection.
+
+Generate one 16:9 horizontal realistic cinematic documentary still for the `在路上` MV style anchor C: "muddy starlight".
+
+The camera is extremely low, near the muddy ground. The frame is dominated by an irregular muddy puddle in the foreground and middle ground, surrounded by heavy wet mud, small stones, broken shoe prints, dark road grit, and uneven puddle edges. The puddle reflects a dark night or deep pre-dawn sky. The reflected stars are visible and broken: scattered small cold points, slightly fragmented by ripples, dirt, uneven water, and mud. The reflection should feel discovered inside the dirty water, not like a fantasy portal.
+
+The physical world remains real: sticky wet mud, dull earth-brown sludge, coal-black road edges, cold blue-gray night, faint silver-blue sky reflection, low contrast, restrained detail. The image should feel like poetic realism, not fantasy.
+
+A small old walking boot or the edge of the traveler's lower body may appear only at the far left or lower edge of the frame, partially cropped, under about 15% of the image. It should confirm bodily presence and road cost, not become a shoe advertisement or a character portrait. The puddle and broken starlight remain the subject.
+
+Lighting:
+
+- dark night or deep pre-dawn;
+- very low ambient light;
+- cold silver-blue star reflection;
+- no broad warm horizon;
+- no visible sun;
+- no sunset glow;
+- no sunrise glow;
+- no large cinematic rim light;
+- dark areas retain depth without becoming flat black.
+
+Composition:
+
+- 16:9 horizontal;
+- mud and puddle fill most of the frame;
+- horizon, if visible, is minimal and dark;
+- no face;
+- no hero body;
+- no centered person;
+- no clean landscape beauty shot;
+- no text, no labels, no logo, no UI.
+
+Mood:
+
+- quiet;
+- low;
+- heavy;
+- restrained;
+- lonely but not hopeless;
+- the feeling of "low place sees high place";
+- continuation, not victory.
+
+Strictly avoid:
+
+- sunrise;
+- sunset;
+- bright dawn;
+- broad orange / gold horizon light;
+- pretty tourism reflection;
+- fantasy magic;
+- glowing portal;
+- cosmic explosion;
+- surreal galaxy water that overwhelms the mud;
+- clean mirror-like lake;
+- beautiful landscape postcard;
+- shoe advertisement;
+- boot closeup as product shot;
+- person as main subject;
+- camping or firelight;
+- neon / cyberpunk / sci-fi;
+- UI, code, symbols, text, labels, watermark;
+- only mud and no stars;
+- only stars and no mud.
+
+The final image should make the viewer first see real mud and a muddy puddle, then discover broken starlight inside the puddle.
+
+## Optional Reference Handling At Generation Time
+
+If the host image tool supports image references:
+
+- attach `intake/imgs/C.泥泞里的星空蓝图.png`;
+- label it as problem reference / composition hint;
+- explicitly instruct the model to darken and rebuild the scene, not copy the warm sky.
+
+If the host image tool does not support image references:
+
+- use the rendered prompt as a standalone prompt;
+- preserve the same problem-reference analysis in the review file.
+
+## Acceptance Criteria
+
+C repair passes only if:
+
+1. The first impression is muddy puddle and real wet mud.
+2. Broken star reflection is clearly visible in the puddle.
+3. The image reads as night or deep pre-dawn, not sunrise or sunset.
+4. Warm horizon light is absent or extremely minimal.
+5. The boot / traveler element, if present, stays peripheral and under about 15%.
+6. The result feels documentary-real and restrained, not magical.
+7. The image can become the C sample in the final style anchor board.
+
+## Failure Conditions
+
+Reject the generation if any of these happen:
+
+- looks like sunrise or sunset;
+- puddle becomes a beautiful warm landscape reflection;
+- stars are missing or too weak;
+- mud becomes too clean;
+- image becomes fantasy / cosmic magic;
+- boot becomes a shoe ad;
+- traveler becomes the main subject;
+- visual style no longer fits A/B/D.
+
+## Review File To Create After Generation
+
+After user approval and generation, created:
+
+```text
+visual-system/style/style-anchor-c-repair-review-v1.md
+```
+
+That review compares the generated C repair against:
+
+- current C problem reference;
+- A/B/D style continuity;
+- the acceptance criteria above.
diff --git a/projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-review-v1.md b/projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-review-v1.md
new file mode 100644
index 0000000..957fb71
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-review-v1.md
@@ -0,0 +1,110 @@
+# Style Anchor C Repair Review V1
+
+## Status
+
+- Document status: user accepted C repair candidate.
+- Project: `在路上`.
+- Anchor: C｜泥泞里的星空.
+- Generated image: `visual-system/style/style-anchor-c-repair-v1.png`.
+- Prompt source: `visual-system/style/style-anchor-c-repair-prompt-v1.md`.
+- Execution mode: `gpt-image-2` Skill Mode B, host image-generation tool.
+- Review result: accepted as the C sample for the common style anchor board.
+
+## Generation Trace
+
+Mode check command:
+
+```powershell
+node C:\Users\wangq\.agents\skills\gpt-image-2\scripts\check-mode.js --json
+```
+
+Mode check 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)"
+}
+```
+
+Generation path:
+
+- Garden local generation was not enabled.
+- The prompt was rendered through the `gpt-image-2` Skill workflow.
+- The approved rendered prompt was sent to the host image-generation tool.
+- The generated output was copied into the project.
+
+Default host output:
+
+```text
+C:\Users\wangq\.codex\generated_images\019ef2f7-e630-7303-a0e8-e5e7a86912fc\ig_07b004f488501211016a3a4256b1388196bdde8e5513f400f5.png
+```
+
+Project copy:
+
+```text
+visual-system/style/style-anchor-c-repair-v1.png
+```
+
+## Inspection Method
+
+- Local visual inspection was performed against the acceptance criteria.
+- External image-analysis upload was not used; the attempted external vision analysis was rejected because it would send a project image to a third-party service without explicit user authorization.
+
+## Acceptance Criteria Review
+
+| Criterion | Assessment | Notes |
+| --- | --- | --- |
+| First impression is muddy puddle and real wet mud | pass | The frame is dominated by wet mud, puddle edges, dark grit, stones, and dirty water. |
+| Broken star reflection is clearly visible | pass | Star points are visible in the puddle and are interrupted by mud, stones, and uneven water. |
+| Reads as night or deep pre-dawn | pass | The palette is cold blue-gray and dark; it does not read as sunrise or sunset. |
+| Warm horizon absent or minimal | pass | No broad orange/gold horizon is visible. |
+| Boot/traveler stays peripheral under about 15% | pass | The boot is cropped at the far left and remains secondary to the puddle. |
+| Documentary-real and restrained, not magical | mostly pass | The mud is real and grounded; the star reflection is slightly luminous but still acceptable as poetic realism. |
+| Can become the C sample in the final board | pass | User accepted this candidate as the new C blueprint. |
+
+## Comparison To Previous C Problem Reference
+
+The generated candidate fixes the main known problem in the previous C blueprint:
+
+- no sunrise or sunset feeling;
+- no broad warm horizon glow;
+- puddle and mud are primary, not a pretty landscape reflection;
+- star reflection is clearer and more central;
+- traveler trace remains peripheral.
+
+Remaining watch point:
+
+- The reflected stars are clean and bright in the lower-right puddle area. This is visually effective, but if the board needs a harsher documentary tone, a later version could make the stars fewer, smaller, and more broken.
+
+## A/B/D Continuity Check
+
+This candidate should remain compatible with A/B/D if the final board keeps:
+
+- low saturation;
+- documentary texture;
+- environment-first composition;
+- small human presence;
+- warmth reserved for D only.
+
+The current C candidate does not introduce neon, cyber, fantasy portal, camping fire, or hero-person framing.
+
+## Decision
+
+User accepted:
+
+```text
+visual-system/style/style-anchor-c-repair-v1.png
+```
+
+Next step:
+
+```text
+visual-system/style/style-anchor-board-spec-v1.md
+```
+
diff --git a/projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png b/projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png
new file mode 100644
index 0000000..ab47e02
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/style/style-anchor-c-repair-v1.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/style/style-anchor-trial-plan.md b/projects/2026-06-23-在路上/visual-system/style/style-anchor-trial-plan.md
new file mode 100644
index 0000000..cb4edba
--- /dev/null
+++ b/projects/2026-06-23-在路上/visual-system/style/style-anchor-trial-plan.md
@@ -0,0 +1,191 @@
+# Style Anchor Trial Plan
+
+## Status
+
+- Document status: user accepted; superseded by style brief/spec/review artifacts.
+- Execution status: planning completed; do not generate images from this file.
+- Scope: common style anchor board for `在路上`.
+- Next gated action after approval: completed through `style-anchor-brief-v1.md`, `style-anchor-board-spec-v1.md`, and `style-anchor-board-review-v1.md`.
+
+## Purpose
+
+The character anchor answers:
+
+> Is this still the same traveler?
+
+The style anchor answers:
+
+> Do these images still belong to the same MV?
+
+This trial should turn the four existing blueprint images into an execution-ready style system without mixing them with the character-anchor workflow.
+
+## Source Images
+
+Blueprint images:
+
+| Anchor | Source file | Current role |
+| --- | --- | --- |
+| A | `intake/imgs/A.风景主导型蓝图.png` | Accepted base for landscape-led road composition. |
+| B | `intake/imgs/B.高位观察型蓝图.png` | Accepted base for high-angle city-edge observation. |
+| C | `intake/imgs/C.泥泞里的星空蓝图.png` | Problem reference; must be rebuilt darker with clearer broken starlight. |
+| D | `intake/imgs/D.夜路篝火型蓝图.png` | Accepted base for restrained night-road fire warmth. |
+
+Local contact sheet:
+
+```text
+visual-system/style/style-blueprints-contactsheet.png
+```
+
+## Source Files Actually Used
+
+Primary project sources:
+
+- `visual-system/visual-system.md`
+- `intake/阶段四-通用表达与视觉系统设计.md`
+- `intake/阶段四-风格锚点板页面结构建议稿.md`
+- `project.md`
+- `execution-plan.md`
+
+## Initial Blueprint Assessment
+
+### A: 风景主导型
+
+Usable as a base anchor.
+
+What works:
+
+- long road;
+- large world / small person;
+- cold landscape-led road-film composition;
+- low-saturation earth and gray palette.
+
+Watch risk:
+
+- light is somewhat sunrise-like. It is acceptable for A, but should not infect C.
+
+### B: 高位观察型
+
+Usable as a base anchor.
+
+What works:
+
+- city edge rather than city promotion;
+- wet ground and rough terrain;
+- person small enough;
+- documentary observation feeling.
+
+Watch risk:
+
+- keep it away from clean skyline / tech-city / cyber UI drift in later prompts.
+
+### C: 泥泞里的星空
+
+Do not use as final C anchor.
+
+What works:
+
+- mud and puddle reality are present;
+- low-angle puddle composition is directionally right.
+
+Problem:
+
+- too close to sunrise / sunset reflection;
+- too warm and bright;
+- broken star reflection is not clear enough;
+- the spiritual function is not yet strong enough.
+
+Required repair:
+
+- dark night or pre-dawn;
+- real muddy puddle as main subject;
+- visible broken star reflection;
+- low-place-sees-high-place feeling;
+- no sunrise, no sunset, no fantasy magic, no shoe advertisement.
+
+### D: 夜路篝火型
+
+Usable as a base anchor.
+
+What works:
+
+- dark road;
+- small fire;
+- restrained warmth;
+- distant faint lights;
+- not a camping party or victory scene.
+
+Watch risk:
+
+- later D prompts must preserve solitude and distance, not become a gathering.
+
+## Recommended Advancement
+
+Do not directly make the final style board by simply placing the four current blueprints together.
+
+Recommended sequence:
+
+1. **Style brief**
+   - Create `style-anchor-brief-v1.md`.
+   - Record A/B/D as usable base references.
+   - Record C as a repair target, not an accepted anchor.
+   - Define the visual rules that the final board must enforce.
+
+2. **C repair prompt/spec**
+   - Create a Skill-based prompt document for C only.
+   - Use `gpt-image-2` workflow and record mode, template/reference, source files, and input image roles.
+   - Generate only after user approval.
+
+3. **C repair generation/review**
+   - Generate one C repair candidate.
+   - Review it against the dark muddy starlight criteria.
+   - If C fails, repair C before moving on.
+
+4. **Style anchor board assembly**
+   - Use accepted A, B, repaired C, and accepted D.
+   - Prefer a deterministic board/layout artifact for readable labels and rules.
+   - Do not rely on image generation to render exact Chinese labels.
+
+5. **Five-shot test package**
+   - Use the accepted character slices and style anchors in `S-01`, `S-10`, `S-46`, `S-47`, `S-51`.
+
+## Board Construction Recommendation
+
+For the final common style anchor board, prefer a layout/composition artifact rather than a single AI-generated board.
+
+Reason:
+
+- A/B/D already exist as useful references.
+- C needs one targeted repair.
+- The board needs exact Chinese labels and rules.
+- Generated text inside an image is not reliable enough to carry execution rules.
+
+Recommended board form:
+
+- one 16:9 dark-gray board;
+- four visual sample cells A/B/C/D;
+- right-side compact rule column: color, light, composition, texture, drift avoidance;
+- bottom review rule line;
+- exact text handled by layout/document layer, not by image generation.
+
+## Candidate Reusable Workflow Notes
+
+Do not promote these as stable Video Workbench knowledge yet, but this trial can test:
+
+1. whether style anchors should be split into accepted base anchors plus repair targets;
+2. whether exact labels should be assembled outside image generation;
+3. whether C-style soul images should always get a separate repair gate;
+4. whether five-shot tests should only start after C is repaired.
+
+## Next Gated Action
+
+After this plan is approved, create:
+
+```text
+visual-system/style/style-anchor-brief-v1.md
+```
+
+That brief should be the project-facing source for:
+
+- C repair prompt/spec;
+- final style anchor board structure;
+- later five-shot style references.
diff --git a/projects/2026-06-23-在路上/visual-system/style/style-blueprints-contactsheet.png b/projects/2026-06-23-在路上/visual-system/style/style-blueprints-contactsheet.png
new file mode 100644
index 0000000..6a30b2d
Binary files /dev/null and b/projects/2026-06-23-在路上/visual-system/style/style-blueprints-contactsheet.png differ
diff --git a/projects/2026-06-23-在路上/visual-system/visual-system.md b/projects/2026-06-23-在路上/visual-system/visual-system.md
index 5cebbe7..86b2aa0 100644
--- a/projects/2026-06-23-在路上/visual-system/visual-system.md
+++ b/projects/2026-06-23-在路上/visual-system/visual-system.md
@@ -105,7 +105,8 @@ Core composition rule:
 ### C: Muddy Starlight
 
 - Function: soul image.
-- Status: priority repair; prior direction had reality but not enough visible starlight.
+- Status: repaired sample accepted.
+- Accepted sample: `style/style-anchor-c-repair-v1.png`.
 - Must show: dark night or pre-dawn muddy puddle, real mud, visible broken star reflection, low-place-sees-high-place feeling.
 - Avoid: sunrise, sunset, pretty reflection only, fantasy magic, shoe advertisement, only mud/no stars, only stars/no mud.
 
@@ -198,7 +199,10 @@ Every generated image should answer:
 | Date | Target | Result | Notes |
 | --- | --- | --- | --- |
 | 2026-06-23 | Extract visual system | initialized | Created from cleaned intake. No images generated yet. |
+| 2026-06-23 | C style retest | accepted | `visual-system/style/style-anchor-c-repair-v1.png` accepted as C sample. |
+| 2026-06-23 | Common style anchor board | accepted | `visual-system/style/style-anchor-board-v1.png` accepted as common style anchor board. |
+| 2026-06-23 | Five-shot visual trial | accepted | `S-01`, `S-10`, `S-46`, `S-47`, and `S-51` generated and accepted by user. |
 
 ## Next Action
 
-Create the character anchor board prompt/spec first, then generate and review the first board.
+Decide the lightweight broad-production flow for the remaining non-generated shots.