Update bundled Repository Governance snapshot#48
Conversation
Assisted-by: OpenAI Codex (GPT-5, autonomous)
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 958bc06388
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| if child.is_dir(): | ||
| dirs.append(child) | ||
| next_dirs.append(child) |
There was a problem hiding this comment.
With a repository-local symlinked directory, for example loop -> .. or a linked package outside the repo, Path.is_dir() follows the symlink and project_walk() pushes it onto the traversal stack. Because every bounded evidence helper calls project_files()/project_dirs() before applying depth limits, the generator can recurse outside the project or hang on symlink cycles instead of producing the governance projection; skip symlinked directories or verify resolved children remain under root before appending them.
Useful? React with 👍 / 👎.
| "docs/*agent*governance*.md", | ||
| "docs/**/*agent*governance*.md", | ||
| ] | ||
| CODE_STYLE_SSOT_FILES = [ |
There was a problem hiding this comment.
When the only code-style SSOT is an explicit tool config such as .prettierrc, eslint.config.js, ruff.toml, mypy.ini, or style settings in pyproject.toml, this new whitelist makes code_style_ssot_refs() report Code Style as missing. The generated projection then tells agents to ask for NEEDS_CLARIFICATION:CODE_STYLE instead of reading the actual formatter/linter config, so include the existing CODE_STYLE_FILES entries, or equivalent parsed pyproject.toml refs, in the index source refs.
Useful? React with 👍 / 👎.
| relative = path.relative_to(root) | ||
| if len(relative.parts) - 1 > max_parent_depth: | ||
| continue | ||
| if any(relative.match(pattern) for pattern in patterns): |
There was a problem hiding this comment.
For SSOT docs nested more than one directory under docs/ or adr/, for example docs/platform/backend/architecture-overview.md, Path.match('docs/**/architecture*.md') does not match recursively the way these patterns imply, so bounded_glob_files() misses valid refs even within EVIDENCE_SCAN_MAX_PARENT_DEPTH. Use root.glob or fnmatch with explicit depth handling so the ** entries actually cover the documented nested SSOT locations.
Useful? React with 👍 / 👎.
Allow fork-hosted community integration PRs to reuse the pr-template route without requiring upstream maintainer approval, while preserving the approval requirement for github/spec-kit. Assisted-by: OpenAI Codex (GPT-5, autonomous)
|
Posted on behalf of @bigsmartben by OpenAI Codex (GPT-5). Updated the community integration validation workflow in commit 8ef2ab9. The workflow now passes the current repository name into Validation run locally:
GitHub Actions status after push: |
1 participant
Description
Update the bundled
repository-governanceextension snapshot frombigsmartben/spec-kit-agent-governancesource commit42f0cb04891a29a4c05973b7fa5e746a7e0f8dd4.This pulls in the SSOT index projection changes, explicit Zed target mapping, bounded repository evidence scanning, narrower generated target content, and matching bundled README/changelog updates. The bundled install URL now points at the exact source commit archive used for this snapshot instead of an older tag.
Testing
uv run specify --helpuv sync && uv run pytestValidation performed:
uv run python scripts/community/validate_integration.py --repo-root . --branch community/repository-governance-ssot-index— passed.\.venv\Scripts\python -m py_compile extensions\repository-governance\scripts\generate_repository_governance.py— passed.\.venv\Scripts\python -m pytest tests\integrations\test_cli.py::TestGitExtensionOptIn::test_community_extensions_and_workflow_preset_auto_installed tests\integrations\test_cli.py::TestGitExtensionOptIn::test_no_git_keeps_community_defaults tests\test_arch_templates.py tests\integrations\community_defaults.py— 8 passed.\.venv\Scripts\python -m pytest— attempted, timed out after 10 minutes and ended withOSError: [Errno 22] Invalid argumentwhile flushing stdout, so no full-suite pass is claimedCommunity Catalog Submission
AI Disclosure
Prepared by OpenAI Codex (GPT-5) on behalf of @bigsmartben. Codex synchronized the bundled extension snapshot, updated metadata, ran validation, committed, pushed, and opened this PR under direct user instruction.