How real projects brief Codex, Cursor and every other agent that reads AGENTS.md.
MistralAGENTS.mdmistralai/mistral-vibe/vibe/cli-rust/AGENTS.md
paulvezia/rewrite_rust branch, always create a worktree and fix the code there - No unit tests allowed in rust files !! ## First MVP - We need basic conversation working, same UI, all other features
WaishnavAGENTS.mdWaishnav/devspace/AGENTS.md
workflow to file the PR and `babysit-pr` to monitor CI and reviews. ## Where code lives - `src/server.ts` — MCP server setup, tool registration, and response wiring. - `src/workspaces.ts` — workspace lifecycle, instructions, skills
NVIDIAAGENTS.mdNVIDIA/Model-Optimizer/AGENTS.md
feedback - **Judge each comment on its merits before acting.** Check it against the current code — reviewers comment on stale diffs, and bot findings (CodeRabbit, Claude) are claims to verify
qixing-jkAGENTS.mdqixing-jk/all-api-hub/AGENTS.md
nvmrc`, and hooks. ## Project boundaries - Keep `src/entrypoints/**` thin; feature UI belongs in `src/features/**`. Shared code must not import options-page entrypoints. Keep React state ownership outside `src/services
AWSAGENTS.mdawslabs/aidlc-workflows/AGENTS.md
recording, exact stage-rule delivery, forwarding-loop enforcement, reviewer read-scope enforcement, review-receipt write-freeze enforcement, code-generation plan-approval enforcement, direct state-transition enforcement, and token-usage folding
homeassistant-aiAGENTS.mdhomeassistant-ai/ha-mcp/AGENTS.md
policy and controls](docs/agents/issue-intake.md) for source handling, maintainer overrides, permissions, and bench validation. ### Automated Code Review After every push, inspect human and automated feedback. Codex and CodeRabbit findings are hypotheses
homeassistant-aiAGENTS.mdhomeassistant-ai/ha-mcp/homeassistant-addon-webhook-proxy/AGENTS.md
transform (token rename + flavor-banner swap) and follows the dev-first flow like code; `CHANGELOG.md` stays per-flavor and gets a manual entry on each promote PR. ## Mutual exclusion Both
homeassistant-aiAGENTS.mdhomeassistant-ai/ha-mcp/tests/AGENTS.md
Test Design Rules Every test under `tests/` follows these. Reviewers check new and changed tests against them. 1. **Test our code, not the platform.** Assert on behaviour this repository owns
MicrosoftAGENTS.mdmicrosoft/fluentui-blazor/AGENTS.md
package version management - `.github/CODEOWNERS` - Code ownership (@vnbaaij @dvoituron) - `docs/contributing.md` - Contribution guidelines - `docs/unit-tests.md` - Detailed unit testing documentation ## Security Considerations - Do not commit sensitive data or credentials - Review the `SECURITY.md` file
callstackAGENTS.mdcallstack/agent-device/AGENTS.md
guards that reconstruct another source of truth. - Keep control-flow narration and review history out of implementation comments. Encode invariants in names, types, boundaries, and tests; reserve comments for public
Agenta-AIAGENTS.mdAgenta-AI/agenta/web/AGENTS.md
components in `web/storybook` (`stories/entity-ui/` for the data-connected `@agenta/entity-ui` surfaces). Cover the states a reviewer cannot reach by clicking — empty, error, read-only, migration. Nothing type-checks that package
thunderbirdAGENTS.mdthunderbird/thunderbolt/AGENTS.md
file components only ### `useEffect` Discipline **Treat every `useEffect` as a code smell until proven necessary.** Before writing or reviewing a `useEffect`, consult https://react.dev/learn/you-might-not-need-an-effect and verify it doesn
jakubkrehelAGENTS.mdjakubkrehel/skills/AGENTS.md
verification checks and the format for a standalone review. See below. Supporting `.md` files carry depth beyond the principle statements: recipes, code patterns, lookup tables. Link each one from
mvanhornAGENTS.mdmvanhorn/cli-printing-press/AGENTS.md
seeing the blocker. Plan documents stay local (see "Plan documents stay local" below). ## Automated code review with Greptile Every PR gets automated Greptile review alongside CI. Resolve every Greptile finding
MicrosoftAGENTS.mdmicrosoft/WindowsAppSDK/AGENTS.md
Clarification - Ambiguous requirements after reviewing docs - Cross-feature impact unclear - API usage patterns not well documented ## Detailed Documentation - [Coding Guidelines](docs/Coding-Guidelines.md) – Complete coding standards and checklist - [Contributor Guide](docs/contributor-guide.md) – Contribution
JetBrainsAGENTS.mdJetBrains/koog/AGENTS.md
Quality Gates Read and follow the Quality Gates section in /TESTING.md before considering any code change complete. ## Architecture ### Core Framework Components **AIAgent** — Main orchestrator that executes strategies in coroutine scopes
gptmeAGENTS.mdgptme/gptme/AGENTS.md
only the CC root; pass `--scope cc` when the memory must appear in Claude Code without recall. ### Inspect and manage ```bash gptme-util memory roots # Which roots are active
crmneAGENTS.mdcrmne/ruby_llm/AGENTS.md
Ground rules - New features need an approved GitHub issue before any code. PRs without one get closed without review. - Never edit generated files: `lib/ruby_llm/models.json`, `lib/ruby_llm/aliases.json`, `docs/_reference/available-models.md`. `rake models` regenerates them
ElasticAGENTS.mdelastic/elasticsearch-py/AGENTS.md
generate a contribution is not contributing. PRs opened this way will be closed without review. If a human asks you to "find something to work on" or points
inkeepAGENTS.mdinkeep/open-knowledge/AGENTS.md
break`, `WARN: a sibling that silently drifts if this changes`, `UPSTREAM(electron/electron#19920): a code shape forced on us from outside this repo`. - **A validated citation**, as in `precedent
An open format for instructions to coding agents, read by Codex, Cursor and others. Think of it as a README written for agents.
At the repository root, with more specific files in subdirectories. Agents read the one closest to the file they're editing.
Setup and test commands, code style, and the rules a new contributor would need to know.
Claude Code reads CLAUDE.md. A one-line CLAUDE.md that points at AGENTS.md covers both.
