agentleFS
Sign inStart free
jlrrequinajlrrequina/pi-project-profile/AGENTS.md

pi-project-profile

AGENTS.md0 starsChanged 45 days agoMITView on GitHub

A π (pi-coding-agent) extension package. README.md is the user-facing reference; this file is for contributors and agents. Add a unit test for every detector or planner change. Add changelog entries under ## Unreleased in CHANGELOG.md; releases are cut by the Release workflow (see CONTRIBUTING.md), never by hand. Workflows pin actions to commit SHAs and must pass actionlint and zizmor.

  • 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

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    No runtime dependencies: Import only from node:*, @earendil-works/pi-coding-agent, @earendil-works/pi-ai (Type), @earendil-works/pi-tui …

    line 8
  2. Very specific:

    Never write into the target repository: All state goes under ~/.pi/agent/project-profile/ or $TMPDIR/pi-project-profile/.

    line 10
  3. Very specific:

    Add a unit test for every detector or planner change.

    line 25

≈ “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

Working on pi-project-profile

A π (pi-coding-agent) extension package. README.md is the user-facing reference; this file is for contributors and agents.

Hard constraints — do not regress

  • Erasable TypeScript only. π loads .ts directly (Node type stripping / jiti): no enum, namespace, parameter properties, or decorators. Use import type, explicit .ts import extensions, "type": "module". tsc enforces erasableSyntaxOnly and verbatimModuleSyntax.
  • No runtime dependencies. Import only from node:*, @earendil-works/pi-coding-agent, @earendil-works/pi-ai (Type), @earendil-works/pi-tui (Box/Text/Markdown).
  • The prompt section must be deterministic per session (no timestamps, branch names, or ordering that depends on iteration order) so prompt caching keeps working.
  • Never write into the target repository. All state goes under ~/.pi/agent/project-profile/ or $TMPDIR/pi-project-profile/.
  • Never auto-run installs, migrations, deploys, or anything network-bound. Tests and builds are confirm-once tiers (TIER_POLICY in types.ts). When unsure, fail to a no-op — never to a wrong action.
  • Bump DETECTOR_VERSION in detect/index.ts whenever detector output shape or semantics change (it invalidates cached profiles).
  • Discoverability metadata is tested (test/package.test.ts). The npm description must stay ≤ 250 characters (the registry truncates at 255; the π catalog card shows ~250). skills/project-profile/SKILL.md follows the Agent Skills spec (name = directory name, description ≤ 1024 characters, states when to use it) and must document every /profile subcommand and config key that exists. llms.txt links only to files that exist. docs/site/index.html, sitemap.xml and .github/workflows/pages.yml must agree (the test checks). When a command, config key or behaviour changes, update the README, the skill and the site together.

Verify your changes

npm run check                          # tsc, must be clean
npm test                               # node --test test/*.test.ts, all passing
node scripts/scan.ts <repo> --checks   # detector output for any real repository
npm run corpus                         # detection on ~20 real repositories (clones them; CI runs it too)
PI_PROJECT_PROFILE_DEBUG=1 pi -p "…"   # live gate trace in $TMPDIR/pi-project-profile/debug.jsonl

Add a unit test for every detector or planner change. Add changelog entries under ## Unreleased in CHANGELOG.md; releases are cut by the Release workflow (see CONTRIBUTING.md), never by hand. Workflows pin actions to commit SHAs and must pass actionlint and zizmor.

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