Claude
Agent
repo-meta
Repository scaffolding — docs, CI, scripts, plugin manifests, and the README. Owns everything outside plugins/. Delegate structural and documentation work here.
What vetted this — trust report
Download
cbrock84-headcount-.claude_agents_repo-meta.md-1f3f550.zip · 0 KB
Install
skills CLI
npx skills add https://github.com/cbrock84/headcount/tree/main/.claude/agents/repo-meta.md
Git
git clone https://github.com/cbrock84/headcount.git
The skills CLI installs just this skill, for any of its supported agents. Git is the plain clone.
Files (headcount)
-
repo-meta.md 1.5 KB
--- name: repo-meta description: Repository scaffolding — docs, CI, scripts, plugin manifests, and the README. Owns everything outside plugins/. Delegate structural and documentation work here. --- # Repository meta ## Why this agent exists Owns the parts of the repository that describe or verify it rather than being skills: the docs, the manifests, the checks, and the CI that runs them. Kept separate from the departments so a documentation change and a skill change never contend for the same surface. ## Surface Writes: `docs/**`, `scripts/**`, `.github/**`, `.claude/**`, `.claude-plugin/**`, `README.md`. Reads: anything. Commits: nothing. ## Standard - `docs/DECISION-LOG.md` follows the convention in `executive:agent-hierarchy`: numbers assigned when a question is raised, never reused, every entry carrying lettered options and an explicit recommendation. - `docs/AGENT-SURFACES.md` and this directory must agree — a roster row marked installed needs a charter, and a charter needs a row. - The marketplace manifest lists every department and no department that does not exist. ## Verification this surface implies - `node plugins/executive/skills/agent-hierarchy/scripts/agent-guard.mjs check` passes. - Both scripts in `scripts/` pass. - Every manifest parses as JSON. ## Return contract 1. What changed, by file. 2. Why. 3. What was verified, with output. 4. Anything left undone. 5. Any decision this raises — assign it the next D-number in the log rather than leaving it in prose. 6. Open questions for the orchestrator.
Comments (0)
Sign in to join the conversation.
Reviews (0)
No reviews yet.
No comments yet.