AGENTS.md examples from real projects

basecampAGENTS.md

fizzy

basecamp/fizzy/AGENTS.md

disagrees, take the better path and flag the conflict; invariants (data loss, security, CI gates) are surfaced, not overridden. Attack your own diff before calling it done. ## Deploy Default branch

No other project uses it yet

apacheAGENTS.md

tomcat

apache/tomcat/AGENTS.md

Agent guidance This file is read by automated agents (security scanners, code analyzers, AI assistants) operating on this repository. It points them at the human-authored references they should consult

No other project uses it yet

hatchet-devAGENTS.md

hatchet

hatchet-dev/hatchet/AGENTS.md

Hatchet server instance (engine, API, `hatchet-lite`, docker-compose, or helm) must set `SERVER_SECURITY_CHECK_ENABLED=false`. The check defaults to enabled and phones home to `security.hatchet.run`; CI must

No other project uses it yet

trailofbitsAGENTS.md

skills

trailofbits/skills/AGENTS.md

plugin](https://github.com/EveryInc/compound-engineering-plugin) - Production plugin structure - [getsentry/skills](https://github.com/getsentry/skills) — Production Sentry skills; `security-review` is a standout routing + progressive disclosure example **For Claude:** Use the `claude-code-guide

No other project uses it yet

zeekAGENTS.md

zeek

zeek/zeek/AGENTS.md

test passed unless it was actually run; an AI-predicted result is not evidence. ## Security-Sensitive Work Read `SECURITY.md` before handling a suspected vulnerability. Consult the relevant parts of `SECURITY_MODEL.md

No other project uses it yet

MoonshotAIAGENTS.md

kimi-code

MoonshotAI/kimi-code/docs/AGENTS.md

prerequisites) Supplementary context / use-case list (optional) Diagram (optional) ::: warning Banner (deprecation, breaking change, security notice — after opening content, before first ##) ## First section Body… ## Next steps (optional, only when related

No other project uses it yet

podman-desktopAGENTS.md

podman-desktop

podman-desktop/podman-desktop/AGENTS.md

extensions, tray | | Renderer | `packages/renderer` | Svelte 5+ UI in Chromium, IPC to main | | Preload | `packages/preload*` | Security bridge exposing safe APIs | ### Key Packages ``` packages/main/ - Electron main process (Inversify DI, IPC handlers) packages/renderer

No other project uses it yet

agegrAGENTS.md

pi-web

agegr/pi-web/AGENTS.md

components/SelectorRow.tsx`. - [files-and-access.md](docs/agents/files-and-access.md): worktrees and project grouping, the file access allow-list (the `/api/files` security boundary), file tree visibility and its open directories per folder, uploads and chat file drops

nullclawAGENTS.md

nullclaw

nullclaw/nullclaw/AGENTS.md

without justification. - `MaxRSS` during `zig build test` must stay well under 50 MB. 3. **Security-critical surfaces are first-class** - `src/gateway.zig`, `src/security/`, `src/tools/`, `src/runtime.zig` carry high blast radius. - Defaults

logancyangAGENTS.md

obsidian-copilot

logancyang/obsidian-copilot/AGENTS.md

order: 1. **Correctness** — logic errors, data loss, unhandled failure paths, concurrency hazards. 2. **Security** — authorization bypass, injection, secrets or private data leaving the codebase. 3. **Breaking changes and migrations** — changed

permissionlesstechAGENTS.md

bitchat-android

permissionlesstech/bitchat-android/AGENTS.md

structured coroutines and suspend I/O, and never block the main thread. Protocol and security changes must remain fail-closed and cross-client compatible. Update the relevant specification and golden-vector

Rich-HarrisAGENTS.md

degit

Rich-Harris/degit/AGENTS.md

docs/USAGE.md](docs/USAGE.md) | | License text | [LICENSE.md](LICENSE.md) | | Contributing flow, PR checks, commit style | [docs/CONTRIBUTING.md](docs/CONTRIBUTING.md) | | Security policy and reporting | [docs/SECURITY.md](docs/SECURITY.md) | | Community expectations | [docs/CODE_OF_CONDUCT.md](docs/CODE_OF_CONDUCT.md) | | Published CLI help text | [assets/help.md](assets/help.md

No other project uses it yet

leejetAGENTS.md

stable-diffusion.cpp

leejet/stable-diffusion.cpp/AGENTS.md

pitfalls. For general contribution rules, including PR scope, commit conventions, code style, dependency updates, security hygiene, and AI-assisted contribution policy, see `CONTRIBUTING.md`. --- ## Agent Operating Rules Before analyzing or modifying

NirDiamantAGENTS.md

Prompt_Engineering

NirDiamant/Prompt_Engineering/AGENTS.md

engineering, from basic prompt structure through chain-of-thought, self-consistency, task decomposition, prompt security and systematic prompt evaluation. Each notebook explains one technique, implements it, and shows what changes

btcpayserverAGENTS.md

btcpayserver

btcpayserver/btcpayserver/AGENTS.md

facing [maintainer handbook](docs/maintainers/README.md) before changing repository behavior or process. `SECURITY.md` remains the canonical security policy. When running tests, always set `PLAYWRIGHT_HEADLESS=true` so Playwright cannot open browser windows

macvim-devAGENTS.md

macvim

macvim-dev/macvim/AGENTS.md

stability periods — see `runtime/doc/develop.txt` `*design-policy*`. - **During a stability period** only clear bug fixes, security fixes, documentation updates, translations, and runtime file updates are accepted. No new features, no backwards

homeassistant-aiAGENTS.md

ha-mcp

homeassistant-ai/ha-mcp/AGENTS.md

repository is `homeassistant-ai/ha-mcp`; the package is `ha-mcp` on PyPI. ## Security Read [`SECURITY.md`](SECURITY.md) before changing authentication, OAuth, webhooks, network exposure, proxies, filesystem access, or trust boundaries

go-gitAGENTS.md

go-git

go-git/go-git/AGENTS.md

critical fixes target `main`. Backports to `releases/v5.x` must be motivated by a bug or security issue, and agreed with a maintainer. ## Code Conventions - **Go version**: stay within the range declared

No other project uses it yet

DolibarrAGENTS.md

dolibarr

Dolibarr/dolibarr/.agents/AGENTS.md

tags - **dolGetButton***: For consistent button styling - **setEventMessages**: For user feedback messages --- ## Security - Guard page access with `restrictedArea($user, 'module', $id, 'table')` or a specific test that deny access with `accessforbidden

No other project uses it yet

midudevAGENTS.md

preguntas-entrevista-react / vercel-react-best-practices

midudev/preguntas-entrevista-react/.agents/skills/vercel-react-best-practices/AGENTS.md

invoked directly. Next.js documentation explicitly states: "Treat Server Actions with the same security considerations as public-facing API endpoints, and verify if the user is allowed to perform a mutation