Doc Update

Inputs

• input.json with:
  • topic (string): e.g. "mcp", "skills", "governance".
  • extra_targets (array, optional): Additional paths to include in the update set.

Outputs

• Writes actual.json with:
  • topic (string)
  • doc_system (array of strings)
  • recommended_targets (array of strings)
  • notes (array of strings)

Workflow

1. Confirm intent gate (do not edit CANON or existing CONTEXT unless explicitly requested).
2. Use CORTEX/query.py to locate entrypoints; avoid raw filesystem discovery.
3. Update the smallest set of docs that inform all agents (README, CONTEXT/maps/ENTRYPOINTS, MCP docs, relevant ADRs).
4. For MCP updates, include:
   • The recommended entrypoint (LAW/CONTRACTS/ags_mcp_entrypoint.py).
   • The log location (LAW/CONTRACTS/_runs/mcp_logs/).
   • The verification skills (mcp-smoke, mcp-extension-verify).
   • Client config examples for Windows and WSL where relevant.
5. Run python3 TOOLS/critic.py and python3 CONTRACTS/runner.py after changes.

Constraints

• Must not scan the filesystem directly (use CORTEX/query.py for discovery).
• Must not modify canon or context unless explicitly authorized.
• Must not write artifacts outside allowed roots.

Fixtures

• fixtures/basic/

required_canon_version: >=3.0.0