docs: add project supplier boundaries

AGENTS.md

@@ -203,8 +203,37 @@ At minimum, keep these files consistent:
 - `docs/DATA_CONTRACT.md`
 - `docs/WORKFLOW.md`
 - `docs/DECISIONS.md`
+- `PROJECTS.md`
 
-## 13. Decision Logging
+## 13. Cross-Repository Collaboration Rules
+
+This repository is the product and application boundary for The Mindscape of Bro Tsong.
+
+Use neighboring repositories through explicit request channels:
+
+- `requirements/ccpe/` for expert agents, CCPE-Lite prompt cards, Agent Specs, Committee Specs, Runtime Specs, Model Card governance, Model Index governance, invocation protocols, evaluation rubrics, and integration registrations.
+- `requirements/skills-vault/` for reusable automation tools, deterministic scripts, batch processors, validation helpers, extraction utilities, format transformers, and installable automation Skills.
+
+Do not copy canonical CCPE artifacts into this repository.
+
+Do not copy reusable automation Skill source into this repository.
+
+Use installed runnable Skills from:
+
+```text
+C:\Users\wangq\.agents\skills
+```
+
+If a missing capability blocks model extraction:
+
+1. Classify whether it is a CCPE need or a skills-vault need.
+2. Create a request file in the appropriate `requirements/` directory.
+3. Pause the dependent extraction step unless the project owner explicitly says to solve it inside this repository.
+4. Resume only after the required spec, rule, tool, Skill, or owner decision is available.
+
+Third-party frameworks such as CrewAI or LangGraph belong in this repository only when they implement this product. CCPE may define the agent contract, runtime governance, authority, evaluation, and integration registration. This repository owns application adapters, state, deployment, and product behavior.
+
+## 14. Decision Logging
 
 Any structural decision should be recorded in:
 
@@ -219,8 +248,9 @@ Examples:
 - Why QPI and Intellectual Archaeology are the first sample models
 - Why selector is rule-based in v0.1
 - Why `model_library_mvp` is a phase name rather than a nested repository root
+- Why a missing capability was routed to `ccpe-system`, `skills-vault`, or implemented locally
 
-## 14. Definition of Done
+## 15. Definition of Done
 
 A task is done only when:
 
@@ -231,8 +261,9 @@ A task is done only when:
 - Validation status is documented
 - Non-goals have not been violated
 - Any open questions are listed in the handoff document
+- Any missing supplier capability is recorded in `requirements/ccpe/` or `requirements/skills-vault/`
 
-## 15. Handoff Requirement
+## 16. Handoff Requirement
 
 At the end of a work session, create or update:
 

PROJECTS.md

@@ -0,0 +1,126 @@
+# Project Relationships
+
+This file is the local project map for The Mindscape of Bro Tsong. It exists so future Codex sessions know what this repository owns, what it requests from neighboring repositories, and when to pause instead of improvising missing tools.
+
+## Local Identity
+
+`the-mindscape-of-bro-tsong` is the product and application repository for The Mindscape of Bro Tsong.
+
+The current phase is:
+
+```text
+model_library_mvp
+```
+
+This repository owns:
+
+- Product-specific model library data
+- Model JSON files
+- Human-readable model cards used by this product
+- Source article and evidence indexes selected for this product
+- Regression cases for product model stability
+- Selector rules and examples for this product
+- Application implementation when the project later adopts frameworks such as CrewAI, LangGraph, or a server runtime
+- Project-specific requirements, handoffs, validation reports, and decision records
+
+This repository does not own canonical expert-agent specifications, reusable automation Skill source, or upstream source archives.
+
+## Neighboring Projects
+
+| Project | Relationship To This Repository | This Repository Consumes | This Repository Produces |
+| --- | --- | --- | --- |
+| `knowledge-vault` | Upstream source archive and discussion memory | canonical articles, source discussions, prior model material, planning records | selected source references, feedback about missing source organization |
+| `ccpe-system` | Supplier for expert-agent architecture, model governance, runtime specs, and integration registrations | Agent Specs, Runtime Specs, Model Card rules, Model Index rules, invocation protocols, evaluation rubrics, integration registration guidance | capability requests under `requirements/ccpe/`, product feedback from real use |
+| `skills-vault` | Supplier for reusable automation Skill source | installed automation Skills from `.agents/skills`, Skill usage rules, validated scripts or CLIs | capability requests under `requirements/skills-vault/`, reusable automation demand discovered during model extraction |
+| `.agents/skills` | Local runtime installation surface | installed runnable Skills | no source ownership; runtime use only |
+| `writing-workbench` | Related downstream writing workspace | possible writing-oriented usage feedback | model usage or output requirements if writing workflows later depend on this system |
+| `video-workbench` | Related downstream dimensional-output workspace | requirement directory pattern and possible output workflow needs | model usage or output requirements if video workflows later depend on this system |
+
+## Boundary Rules
+
+- Product data, application code, selector behavior, and product-specific validation belong in this repository.
+- Canonical Agent Specs, Runtime Specs, Model Index governance, no-simulation protocols, and integration registrations belong in `ccpe-system`.
+- Reusable automation Skill implementation source belongs in `skills-vault`.
+- Installed runnable Skills are used from `C:\Users\wangq\.agents\skills`, not imported directly from `skills-vault`.
+- Upstream article archives and long discussion records remain in `knowledge-vault`; this repository stores selected source records and evidence references needed by the product.
+- CrewAI, LangGraph, server routes, persistence, queues, deployment scripts, and application monitoring belong in this repository when they implement this product. CCPE may design the agent contract; it should not become the application runtime.
+
+## Request Channels
+
+This repository has two supplier request channels.
+
+### CCPE Requests
+
+Use:
+
+```text
+requirements/ccpe/
+```
+
+Use this channel when the project needs:
+
+- Expert-agent definition
+- CCPE-Lite prompt card
+- Agent Spec
+- Committee Spec
+- Runtime Spec
+- Model Card or Model Index governance
+- Invocation protocol
+- Evaluation rubric
+- Integration registration for a framework, Skill, CLI, API, or installed capability
+
+Do not store canonical CCPE artifacts in this repository. Store the request here, then let `ccpe-system` own the supplied artifact.
+
+### skills-vault Requests
+
+Use:
+
+```text
+requirements/skills-vault/
+```
+
+Use this channel when the project needs:
+
+- Deterministic automation tool
+- Script-backed Skill
+- Batch processor
+- Format conversion helper
+- Validation helper
+- Source splitting or extraction utility
+- Installable local automation Skill
+- Reusable CLI wrapper
+
+Do not store reusable automation Skill source in this repository. Store the request here, then let `skills-vault` own the implementation source if the need becomes stable and reusable.
+
+## Model Extraction Work Mode
+
+Model extraction is a duplex workflow:
+
+```text
+1. Build or request the workflow and supporting tools.
+2. Use the available tools to process model content.
+```
+
+During extraction, if a missing reusable capability is discovered:
+
+1. Classify the missing capability.
+2. If it is agent/spec/runtime/governance work, create a request in `requirements/ccpe/`.
+3. If it is reusable automation/tooling work, create a request in `requirements/skills-vault/`.
+4. Pause the dependent extraction step unless the project owner explicitly says this repository should solve the need directly.
+5. Resume model extraction only after the required tool, Skill, spec, or rule is available, or after the owner changes the scope.
+
+This prevents the model extraction phase from silently turning into uncontrolled framework building or tool development.
+
+## Framework Rule
+
+Third-party agentic frameworks such as CrewAI or LangGraph are allowed only when they serve a concrete product implementation need.
+
+Routing:
+
+```text
+Framework contract, agent role, authority, evaluation, and integration registration -> ccpe-system
+Framework implementation, adapters, state, deployment, and product behavior -> this repository
+Reusable deterministic helpers discovered while implementing -> skills-vault request first
+```
+
+Do not adopt a framework only because it is available. Record the product reason, expected benefit, and boundary decision before introducing it.

README.md

@@ -131,6 +131,8 @@ It is a simple demo that recommends candidate models based on:
 
 This repository consumes selected source material and model definitions from the surrounding ecosystem, but it remains the product/system boundary for The Mindscape of Bro Tsong.
 
+See `PROJECTS.md` for the operative cross-repository boundary map, including request channels for `ccpe-system` and `skills-vault`.
+
 ## 12. Current Status
 
 Initial project setup.

docs/DECISIONS.md

@@ -57,3 +57,29 @@ Reason:
 `model_library_mvp` is the current phase and subsystem goal, not a separate root project.
 
 The repository should not start as `the-mindscape-of-bro-tsong/model_library_mvp/` because that would add unnecessary root-vs-subproject ambiguity before the product needs a monorepo layout.
+
+## Decision 007: Use supplier request channels for CCPE and skills-vault
+
+Status: Accepted
+
+Reason:
+
+This repository is a product and application workspace. It should not silently absorb expert-agent specification work or reusable automation Skill source work.
+
+Requests for agents, runtimes, model governance, invocation protocols, evaluation rules, and integration registrations should be recorded under `requirements/ccpe/`.
+
+Requests for deterministic reusable automation tools and installable Skills should be recorded under `requirements/skills-vault/`.
+
+If a supplier capability blocks model extraction, the dependent extraction step should pause unless the project owner explicitly says to solve the need locally.
+
+## Decision 008: Third-party frameworks belong to product implementation, not CCPE or skills-vault
+
+Status: Accepted
+
+Reason:
+
+Frameworks such as CrewAI and LangGraph may be useful later for implementing product behavior. The concrete application adapters, state, deployment, and product runtime belong in this repository.
+
+CCPE may define the agent contract, runtime governance, authority, evaluation, and integration registration.
+
+skills-vault may provide reusable deterministic helper Skills discovered during implementation.

docs/WORKFLOW.md

@@ -59,3 +59,39 @@ At the end of each work session:
 4. Separate assumptions from verified facts.
 5. List questions that require product or CCRA judgment.
 6. Suggest the smallest useful next tasks.
+
+## 6. Supplier Request Workflow
+
+When this repository needs a capability owned by a neighboring repository:
+
+1. Classify the need.
+2. Use `requirements/ccpe/` for expert-agent, runtime, model-governance, invocation, evaluation, or integration-registration needs.
+3. Use `requirements/skills-vault/` for deterministic automation, reusable scripts, validation helpers, extraction helpers, or installable Skill needs.
+4. Write one request file per missing capability.
+5. Pause the dependent model extraction or implementation step unless the project owner explicitly says to solve the need locally.
+6. Resume after the supplier artifact, installed Skill, or owner decision is available.
+
+## 7. Duplex Model Extraction Workflow
+
+Model extraction has two concurrent tracks:
+
+```text
+Workflow/tooling track -> request or build the process and supporting capabilities.
+Content-processing track -> use available capabilities to extract, validate, and stabilize models.
+```
+
+Do not let the content-processing track silently absorb missing tooling work.
+
+If a tool gap appears, record it as a supplier request and stop the blocked extraction step.
+
+## 8. Framework Adoption Workflow
+
+Third-party agentic frameworks such as CrewAI or LangGraph may be introduced only for concrete product implementation.
+
+Before adoption:
+
+1. Record the product reason.
+2. Identify whether CCPE needs to supply an Agent Spec, Runtime Spec, authority rule, evaluation rule, or integration registration.
+3. Identify whether skills-vault needs to supply reusable deterministic helpers.
+4. Keep framework adapters, state, deployment, and product-specific behavior in this repository.
+5. Record the boundary decision in `docs/DECISIONS.md`.

reports/Codex_工程产物摘要_v0.1.md

@@ -14,6 +14,8 @@ Task: Initialize The Mindscape of Bro Tsong repository for the `model_library_mv
 - Added Codex working rules.
 - Added docs for project brief, data contract, workflow, decisions, non-goals, and handoff template.
 - Added purpose README files for all initial project folders.
+- Added cross-repository relationship rules after reading `ccpe-system` and `skills-vault` documentation.
+- Added supplier request channels for CCPE and skills-vault capability requests.
 
 ## 3. Files Created
 
@@ -34,10 +36,17 @@ Task: Initialize The Mindscape of Bro Tsong repository for the `model_library_mv
 - `scripts/README.md`
 - `reports/README.md`
 - `reports/Codex_工程产物摘要_v0.1.md`
+- `PROJECTS.md`
+- `requirements/ccpe/README.md`
+- `requirements/skills-vault/README.md`
 
 ## 4. Files Modified
 
-- None. This was the initial file creation.
+- `README.md`
+- `AGENTS.md`
+- `docs/WORKFLOW.md`
+- `docs/DECISIONS.md`
+- `reports/Codex_工程产物摘要_v0.1.md`
 
 ## 5. Validation Status
 
@@ -45,6 +54,7 @@ Task: Initialize The Mindscape of Bro Tsong repository for the `model_library_mv
 - Markdown documentation created.
 - No schema, model JSON, selector config, or Python scripts have been implemented yet.
 - No runtime validation is available yet.
+- Cross-repository request channels are documented but do not yet contain concrete capability requests.
 
 ## 6. Assumptions Made
 
@@ -52,11 +62,15 @@ Task: Initialize The Mindscape of Bro Tsong repository for the `model_library_mv
 - `model_library_mvp` is the current phase and subsystem goal, not a nested directory.
 - The first model extraction phase will focus on QPI and Intellectual Archaeology.
 - External source archives remain in `knowledge-vault`; this project stores selected model assets and source references needed for the product.
+- CCPE supplies reusable expert-agent, runtime, model-governance, evaluation, and integration-registration artifacts.
+- skills-vault supplies reusable automation Skill source when a repeated tool need becomes stable.
+- CrewAI, LangGraph, or similar framework implementation belongs in this repository only when needed for this product.
 
 ## 7. Deviations From Plan
 
 - The recommended nested `model_library_mvp/` path from the source initialization brief was intentionally flattened based on the project-root discussion.
 - Schema files and sample model JSON files were not created during initialization because the next phase will perform concrete model extraction.
+- The first initialization commit did not include enough cross-repository boundary detail; this was corrected in a follow-up documentation pass.
 
 ## 8. Known Issues
 
@@ -65,15 +79,18 @@ Task: Initialize The Mindscape of Bro Tsong repository for the `model_library_mv
 - Source article and source excerpt records do not exist yet.
 - Regression cases do not exist yet.
 - Selector rules do not exist yet.
+- No CCPE or skills-vault capability requests have been created yet.
 
 ## 9. Questions for Product / CCRA
 
 - Which source articles should be treated as canonical evidence for QPI?
 - Which source articles should be treated as canonical evidence for Intellectual Archaeology?
 - Should model IDs use English slugs only, or allow bilingual aliases in metadata?
+- During QPI and Intellectual Archaeology extraction, which missing capabilities should pause extraction and become supplier requests?
 
 ## 10. Suggested Next Tasks
 
 1. Define JSON schemas for model specs, source articles, source excerpts, and regression cases.
 2. Extract the QPI model card and model JSON from canonical source material.
 3. Extract the Intellectual Archaeology model card and model JSON from canonical source material.
+4. If extraction exposes missing automation or agentic-governance capabilities, create requests under `requirements/skills-vault/` or `requirements/ccpe/` before continuing the blocked step.

requirements/ccpe/README.md

@@ -0,0 +1,36 @@
+# CCPE Capability Requests
+
+This directory holds supplier tickets for capabilities that should be designed or maintained in:
+
+```text
+C:\Users\wangq\Documents\Codex\ccpe-system
+```
+
+Use this directory when The Mindscape of Bro Tsong needs an expert agent, CCPE-Lite prompt card, Agent Spec, Committee Spec, Runtime Spec, Model Card governance rule, Model Index rule, invocation protocol, evaluation rubric, or external capability registration.
+
+Do not place canonical prompts, agent specs, model cards, runtime specs, or CCPE registrations in this repository. This directory records the request from The Mindscape of Bro Tsong as the client project.
+
+## Request Format
+
+Use one Markdown file per request.
+
+Recommended filename:
+
+```text
+YYYY-MM-DD-short-request-name.md
+```
+
+Each request should include:
+
+- Product context
+- Needed capability
+- Why this belongs in CCPE
+- Current blocked task
+- Input files or source paths
+- Expected output artifact
+- Required human gate
+- Urgency and dependency
+
+## Pause Rule
+
+If model extraction depends on a missing CCPE capability, write the request here and pause the dependent step unless the project owner explicitly says to solve it inside this repository.

requirements/skills-vault/README.md

@@ -0,0 +1,36 @@
+# Skills-Vault Capability Requests
+
+This directory holds supplier tickets for capabilities that should be implemented or maintained in:
+
+```text
+C:\Users\wangq\Documents\Codex\skills-vault
+```
+
+Use this directory when The Mindscape of Bro Tsong needs a deterministic automation tool, script, batch processor, extraction helper, validation helper, export utility, file transformer, or installable automation Skill.
+
+Do not place reusable automation Skill source code in this repository. This directory records the request from The Mindscape of Bro Tsong as the client project.
+
+## Request Format
+
+Use one Markdown file per request.
+
+Recommended filename:
+
+```text
+YYYY-MM-DD-short-request-name.md
+```
+
+Each request should include:
+
+- Product context
+- Repeated action or friction point
+- Input files and expected outputs
+- Safety boundaries
+- Whether file edits are allowed
+- Validation expectation
+- Why this should become a reusable Skill
+- Current blocked task
+
+## Pause Rule
+
+If model extraction depends on a missing reusable automation capability, write the request here and pause the dependent step unless the project owner explicitly says to solve it inside this repository.