pi-harness
AGENTS.md2 starsChanged 21 days agoMITView on GitHub
This is a public pi setup repo. Keep changes generic: no secrets, internal URLs, employer details, or machine-specific paths. The test suite runs on two runtimes: Node for most extensions and Bun for extensions/observatory (it uses bun:sqlite). Use npm run test:node or npm run test:bun to run one half. - npm run check:manifest stays green. It fails when a file under extensions/, rules/, agents/, skills/, or themes/ is neither classified in harness.json nor explicitly excluded. - npm test stays green.…
- Installs packages
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 31,700 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 repository7 other files it gives its agents.
- add-agentSkill
- add-mcp-serverSkill
- add-ruleSkill
- code-reviewSkill
- grill-meSkill
- orchestrateSkill
- pi-upgrade-reviewSkill
The file
We split this AGENTS.md into the rules it gives an agent
9 rules. About 0 are common, 6 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
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
- Distinctive: line 22
Update the README next to the code you changed.
- Very specific: line 24
Use conventional commit subjects: type(scope): imperative description.
- Very specific: line 25
Add a DCO Signed-off-by: trailer (--signoff).
- Distinctive: line 30
Put the code in the right directory.
- Very specific: line 31
Add an entry to harness.json with a tier: core, subagents, jev, integrations, or local.
- Distinctive: line 32
Document it next to the code.
- Distinctive: line 33
Add a test where the behaviour is testable.
- Distinctive: line 38
Prefer graceful degradation when an optional dependency (Jev, Grafana, Bun) is missing.
- Distinctive: line 39
Do not add a dependency that a stranger must pay for to pass CI.
≈ “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.
Files
Working in this repository
This is a public pi setup repo. Keep changes generic: no secrets, internal URLs, employer details, or machine-specific paths.
Setup
npm install
npm run check:manifest
npm testThe test suite runs on two runtimes: Node for most extensions and Bun for extensions/observatory
(it uses bun:sqlite). Use npm run test:node or npm run test:bun to run one half.
Before you push
npm run check:manifeststays green. It fails when a file underextensions/,rules/,agents/,skills/, orthemes/is neither classified inharness.jsonnor explicitly excluded.npm teststays green. If a suite needs a runtime you do not have, say so in the pull request.- Update the README next to the code you changed.
README.mdis the index; directory extensions own their own README. - Use conventional commit subjects:
type(scope): imperative description. - Add a DCO
Signed-off-by:trailer (--signoff). - One concern per pull request.
Adding a component
- Put the code in the right directory.
- Add an entry to
harness.jsonwith a tier:core,subagents,jev,integrations, orlocal. - Document it next to the code.
- Add a test where the behaviour is testable.
Style
- Sparse comments. Do not narrate what the next line does; comment invariants and workarounds.
- Prefer graceful degradation when an optional dependency (Jev, Grafana, Bun) is missing.
- Do not add a dependency that a stranger must pay for to pass CI.
Maintainer notes
The owner's live-sync workflow (sync.sh, publishing from ~/.pi/agent) is documented in
docs/maintainer.md. sync.sh is intentionally untracked.
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
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.

