agentleFS
sergebulaevsergebulaev/car-rental-skills/AGENTS.md

car-rental-skills

AGENTS.md2 starsChanged 23 days agoMITView on GitHub

Read before changing anything. Keep this file and CLAUDE.md aligned. - Single source of truth: .claude-plugin/plugin.json, .claude-plugin/marketplace.json, .codex-plugin/plugin.json, .agents/plugins/marketplace.json. Name, version, author, license, homepage and the bundle description stay identical across them. - .codex-marketplace/car-rental-skills/ is generated. Never edit it by hand: change the root files, then run python3 scripts/sync_codex_marketplace.py. - Default bump is PATCH for every shippable commit. MINOR or MAJOR only when the owner asks. - After a bump: git tag -a vX.Y.Z -m "...", push the tag, then…

  • Commits and pushes

The file

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

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read before changing anything.

    line 3
  2. Distinctive:

    Default bump is PATCH for every shippable commit.

    line 13
  3. Very specific:

    After a bump: git tag -a vX.Y.Z -m "...", push the tag, then gh release create vX.Y.Z --title "vX.Y.Z" --notes "..." --latest (the README …

    line 15
  4. Distinctive:

    Run every check in "Validation" before pushing.

    line 22
  5. Distinctive:

    Never write a market price into a skill or reference as a fact.

    line 51

≈ “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

Agent conventions (Codex) - car-rental-skills

Read before changing anything. Keep this file and CLAUDE.md aligned.

Versioning

  • Single source of truth: .claude-plugin/plugin.json, .claude-plugin/marketplace.json, .codex-plugin/plugin.json, .agents/plugins/marketplace.json. Name, version, author, license, homepage and the bundle description stay identical across them.
  • .codex-marketplace/car-rental-skills/ is generated. Never edit it by hand: change the root files, then run python3 scripts/sync_codex_marketplace.py.
  • Default bump is PATCH for every shippable commit. MINOR or MAJOR only when the owner asks.
  • After a bump: git tag -a vX.Y.Z -m "...", push the tag, then gh release create vX.Y.Z --title "vX.Y.Z" --notes "..." --latest (the README badge reads the Releases API, not tags).

Commits

  • Author must be Sergey: git commit --author="Sergey Bulaev <s@bulaev.org>".
  • Run every check in "Validation" before pushing.

Skill bundle invariants

  • Exactly 8 skills. The count is stated in the plugin manifests, README, root SKILL.md, this file and CLAUDE.md; change all of them together.
  • description: target <= 400 chars, hard <= 510, keyword-rich, a "Not for X (use Y)" sentinel when skills overlap. No em dashes or en dashes in any description:.
  • Skill names are public surface: a rename updates manifests, README table, root SKILL.md and every cross-reference.

Shared references

  • references/catalog-data.md: the market data contract (tools, connector or client, published rates only, sample size and date). Skills cite it instead of repeating it.
  • references/voice-rules.md: the enumerated banned list for customer copy.
  • references/legal-guardrails.md: contract and insurance guardrails.
  • references/untrusted-content.md: data-is-not-instructions. Every skill that reads someone else's text, or calls get_company, search_companies or search_vehicles, carries a ## Untrusted content section citing it; scripts/check_untrusted_rule.py enforces it through its REQUIRED set.

Market data

  • lib/carrentallist_client.py is standard library only and talks to https://mcp.carrentallist.com/mcp. Keep TOOLS in sync with the server: python3 scripts/check_tools.py --live.
  • Never write a market price into a skill or reference as a fact. Examples use the "median $X over N cars, data as of <date>" shape.

Validation

python3 scripts/check_no_secrets.py
python3 scripts/check_frontmatter.py
python3 scripts/check_markdown_references.py
python3 scripts/check_untrusted_rule.py
python3 scripts/check_text_encoding.py
python3 scripts/check_tools.py --live
python3 scripts/sync_codex_marketplace.py
grep -nE '^description:' SKILL.md skills/*/SKILL.md | grep -E '—|–'   # must be empty

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.