agentleFS
Sign inStart free
clouatre-labsclouatre-labs/agentic-coder-skill/AGENTS.md

agentic-coder-skill

AGENTS.md2 starsChanged 2 months agoApache-2.0View on GitHub

The coder skill: a Scout/Guard orchestration skill for AI coding agents, plus its 4 subagents, and the governance githooks that enforce its commit/DCO conventions. Markdown (skill + agent sources) and YAML (per-harness agent frontmatter) — there is no build/test/lint in the traditional sense; the artifacts are prompts and shell hooks, not compiled code. - GPG sign and DCO sign-off: git commit -S --signoff (every commit) - Treat all repositories as public; no secrets, API keys, credentials, or PII - Actions…

  • Reads credentials
  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repository2 other files it gives its agents.

The file

We split this AGENTS.md into the rules it gives an agent

2 rules. About 1 is common, 1 is distinctive, and 0 are very specific to this project.

  • ≈1

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Common:

    Treat all repositories as public.

    line 24
  2. Distinctive:

    Add dependencies without justification in the PR description

    line 63

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

AGENTS.md

Project overview

The coder skill: a Scout/Guard orchestration skill for AI coding agents, plus its 4 subagents, and the governance githooks that enforce its commit/DCO conventions. Markdown (skill + agent sources) and YAML (per-harness agent frontmatter) — there is no build/test/lint in the traditional sense; the artifacts are prompts and shell hooks, not compiled code.

Stack & Commands

# build:  n/a (no compiled artifacts)
# test:   n/a (validated by running the pipeline end-to-end against a real issue)
# lint:   bunx markdownlint-cli2 "**/*.md" (Markdown); shellcheck githooks/* (hooks)
# format: n/a

Development standards

  • GPG sign and DCO sign-off: git commit -S --signoff (every commit)
  • Treat all repositories as public; no secrets, API keys, credentials, or PII
  • Actions pinned to SHA (not tags); actionlint recommended for local workflow validation
  • Training data is stale: verify APIs and versions against installed packages or docs
  • PR labels use the repo's Conventional-Commit-aligned label set: feature, enhancement, bug, documentation, refactor, chore, ci, security (note feature, never the Conventional-Commit type feat)
  • Agent files are generated: edit tools/agents/{pi,claude,goose}-coder-*.yaml (harness frontmatter) and agents-shared/coder-*.md (shared body), then run scripts/generate-coder-agents.sh --write. Never hand-edit agents/{pi,claude,goose}/*; CI fails on drift (--check)
  • Bump the version in skills/coder/SKILL.md for any behavioral change to the pipeline
  • typesafe-ai is an external dependency of the skill: tier classification calls api.typesafe.ai via the judge tool of the decisions-judge MCP server (bin decisions-judge-mcp; source clouatre-labs/decisions-judge-mcp, published on npm) with TYPESAFE_API_KEY from the shell env (never committed, never written to files or handoffs). Every judge path must keep its deterministic inline fallback; this repo documents only the judge's contract (skills/coder/SKILL.md, Constraints #9–#10 and Handoff Validation)

Testing

  • One happy path and one edge case per behavior; no redundant variations
  • AAA pattern (Arrange, Act, Assert); keep each test focused and short
  • There is no automated test suite for the skill/agents themselves; validate changes by running the pipeline against a real GitHub issue in a scratch repo

Visual Aids

  • Tables and code snippets: caption above (*Table N: Description* / *Code Snippet N: Description*)
  • Figures (images and Mermaid diagrams): caption below (*Figure N: Description*)
  • Captions: clear, well written, concise; numbered continuously across the document

Design references

  • README.md — pipeline overview and phase diagram
  • skills/coder/SKILL.md — the orchestrator's own phase-by-phase spec (versioned)

Do not

  • Add dependencies without justification in the PR description
  • Implement features not specified in the assigned issue
  • Modify files outside the scope of the assigned issue
  • Hand-edit generated agent files under agents/pi/, agents/claude/, or agents/goose/

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free