agentleFS
ElasticKnown publisherelastic/go-licence-detector/AGENTS.md

go-licence-detector

AGENTS.md23 starsChanged 43 days agoApache-2.0View on GitHub

See README.md for why this repository exists, who or what uses it, and its responsibility boundary. For every AI tool that materially contributes to code, tests, documentation, configuration, or the substance of a change: For a squash merge, verify that the final squash commit message contains every attribution trailer. GitHub may populate that message from the pull-request description, commit information, or only the pull-request title depending on repository settings, so putting attribution in the PR description improves preservation but does…

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

8 rules. About 2 are common, 4 are distinctive, and 2 are very specific to this project.

  • ≈2

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    Never invent a bot identity, model name, or email address.

    line 20
  2. Distinctive:

    Keep trailers on their own lines without bullets, Markdown emphasis, or surrounding underscores.

    line 21
  3. Distinctive:

    Keep the human author or committer accountable for understanding and verifying the change.

    line 22
  4. Distinctive:

    When preparing a commit or pull request, offer to create it with the correct attribution.

    line 26
  5. Common:

    Never place credentials, tokens, private keys, cookies, or production secret values in tracked files, examples, tests, prompts, logs, or …

    line 30
  6. Very specific:

    Before committing or pushing, scan your work with Gitleaks.

    line 32
  7. Common:

    Treat any detected secret as exposed: stop, remove it, and arrange rotation or revocation before continuing.

    line 34
  8. Very specific:

    Never bypass push protection or secret scanning, select a GitHub bypass reason, use --no-verify, disable a scanner, or add an allowlist or …

    line 35

≈ “Common” is an estimate: at least 1 in 5 of the 36,368 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 guidance

Repository purpose

See README.md for why this repository exists, who or what uses it, and its responsibility boundary.

AI attribution

For every AI tool that materially contributes to code, tests, documentation, configuration, or the substance of a change:

  • Resolve the tool's runtime identity and add one unformatted trailer to the commit message. Use the exact model or agent slug and reasoning effort whenever exposed; omit unavailable components rather than guessing:

    Assisted-by: <tool name> (<most specific verified runtime identity>)
  • Repeat the same trailer in the pull-request description.
  • Preserve the spelling and specificity of exposed runtime values; do not shorten a specific model slug to a broader model family.
  • Preserve valid tool-native attribution, such as Made with [Cursor](https://cursor.com) or a genuine Co-authored-by trailer, in addition to Assisted-by.
  • Never invent a bot identity, model name, or email address.
  • Keep trailers on their own lines without bullets, Markdown emphasis, or surrounding underscores.
  • Keep the human author or committer accountable for understanding and verifying the change.

For a squash merge, verify that the final squash commit message contains every attribution trailer. GitHub may populate that message from the pull-request description, commit information, or only the pull-request title depending on repository settings, so putting attribution in the PR description improves preservation but does not guarantee it.

When preparing a commit or pull request, offer to create it with the correct attribution. If the user will create it manually, show the exact trailers to copy into both places.

Secret handling

  • Never place credentials, tokens, private keys, cookies, or production secret values in tracked files, examples, tests, prompts, logs, or generated output. This repository is public, so anything committed here is world-readable immediately and permanently.
  • GitHub secret scanning and push protection are enabled on this repository. They are the enforcing controls here: push protection blocks a detected secret at push time.
  • Before committing or pushing, scan your work with Gitleaks. It is not pinned in this repository, so use your own installation: gitleaks dir --no-banner --redact=100 .. For an incident check over history, run GIT_CONFIG_GLOBAL=/dev/null gitleaks git --no-banner --redact=100 --log-opts=--all . from a complete clone.
  • The shared elastic/gitleaks-hooks pre-commit hook used elsewhere in Engineering Productivity is not adopted here. That hook lives in a private repository, and a public repository cannot fetch it: fork pull requests receive no secrets, and external contributors have no access. Do not work around this by embedding a token, vendoring the hook, or substituting a different Gitleaks hook.
  • Treat any detected secret as exposed: stop, remove it, and arrange rotation or revocation before continuing. Do not print the value while triaging it. For a public repository, assume a committed secret is compromised the moment it is pushed.
  • Never bypass push protection or secret scanning, select a GitHub bypass reason, use --no-verify, disable a scanner, or add an allowlist or suppression unless the user explicitly requests that exact override after reviewing the finding and consequence.
  • A generic request to finish, commit, push, merge, or open a pull request does not authorize an override.

Download now (ZIP)

  • 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 28,693 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.

Does it work?

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

Reviews can't be read right now.

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.