claude-agent
AGENTS.md5 starsChanged 8 months agoMITView on GitHub
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 repository51 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- apple-notesSkill
- apple-remindersSkill
- bear-notesSkill
- brainstormSkill
- cameraSkill
- content-creatorSkill
- context-healthSkill
- daily-briefingSkill
- deep-readSkill
- draft-messageSkill
- emailSkill
- exa-searchSkill
- gh-issuesSkill
- gif-searchSkill
- github-opsSkill
- google-workspaceSkill
- hue-lightsSkill
- image-genSkill
- imessageSkill
- memory-managerSkill
- migrate-openclawSkill
- notionSkill
- obsidianSkill
- password-managerSkill
- pdf-editorSkill
- peekabooSkill
- placesSkill
- podcast-makerSkill
- quick-researchSkill
- rss-monitorSkill
- security-auditSkill
- session-logsSkill
- setupSkill
- skill-creatorSkill
- slack-opsSkill
- smart-bedSkill
- sonosSkill
- speech-to-textSkill
- spotifySkill
- summarizeSkill
- task-trackerSkill
- text-to-speechSkill
- things-macSkill
- tmux-controlSkill
- trelloSkill
- trend-scoutSkill
- video-extractSkill
- weatherSkill
- whatsappSkill
- x-twitterSkill
The file
We split this AGENTS.md into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is 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.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 7
Run all app commands from app/:
≈ “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
Repository Guidelines
Project Structure & Module Organization
app/ is the main package. Use app/server/ for the Express API, WebSocket handling, agent execution, and route modules; use app/client/ for the React control panel; keep desktop entry code in app/electron/; keep packaging helpers in app/scripts/. Server tests live in app/server/__tests__/. Runtime data lands in memory/ and workspace/. Agent definitions, hooks, and skills live under .claude/, while .mcp.json tracks MCP server configuration. Treat app/dist/ and app/release/ as generated output unless you are intentionally shipping artifacts.
Build, Test, and Development Commands
Run all app commands from app/:
npm install: install dependencies.npm run dev: start the server on127.0.0.1:3456.npm run dev:client: run the Vite client only.npm run dev:all: watch server and client together.npm run build: create the production web bundle.npm test: run the Vitest suite once.npm run test:coverage: run tests with coverage output.npm run electron:devornpm run electron:build:mac|win|linux: run or package the desktop app.
Coding Style & Naming Conventions
Use TypeScript with strict mode and ES modules. Match the surrounding file style; current source uses 2-space indentation, small focused modules, and React function components. Prefer PascalCase for React components such as SettingsPage.tsx, and kebab-case for backend modules such as scheduled-tasks.ts and vector-memory.ts. Keep app/server/index.ts as assembly code; move new domain logic into routes, helpers, or executor modules instead of expanding the entrypoint.
Testing Guidelines
Vitest is configured to discover **/*.test.ts and run in a single fork because tests share a temporary SQLite-backed HOME. Add new backend tests in app/server/__tests__/ or beside the module when that keeps context clear. Name tests by behavior, for example scheduler.test.ts, and cover success paths, failure paths, and persistence side effects. Run npm test before opening a PR; use npm run test:coverage for larger refactors.
Commit & Pull Request Guidelines
Recent history uses Conventional Commits such as feat: and fix:. Keep commit subjects imperative and scoped when useful, for example feat: add goal execution route. Pull requests should include a short summary, linked issue or task, the commands you ran to validate the change, and screenshots for UI updates. Call out edits to .claude/, .mcp.json, secrets handling, or migration scripts explicitly because they affect runtime behavior beyond the web UI.
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.