How real projects brief Codex, Cursor and every other agent that reads AGENTS.md.
tw93AGENTS.mdtw93/Pake/AGENTS.md
pake skill) ├── llms.txt # Agent-facing contract summary (--json, --config, exit codes) ├── scripts/ # Utility scripts ├── tests/ # Unit, integration, and release-flow tests ├── .github/workflows/ # quality/test and release automation ├── default_app_list.json # Popular apps config
rohitg00AGENTS.mdrohitg00/ai-engineering-from-scratch/AGENTS.md
Repo layout ``` phases/ NN-phase-slug/ NN-lesson-slug/ docs/en.md # lesson explainer code/ # implementation + tests quiz.json # 6 questions outputs/ # reusable artifact (skill / prompt / agent / MCP server) README.md # public face; lesson
crewAIIncAGENTS.mdcrewAIInc/crewAI/AGENTS.md
practices and idiomatic patterns. 2. Maintain existing code structure and organization. 3. Write unit tests for new functionality focusing on behaivor and not implementation. 4. Document public APIs and complex
calesthioAGENTS.mdcalesthio/OpenMontage/.agents/skills/flux-best-practices/AGENTS.md
costs at different resolutions. **Creative Exploration / Ideation:** **Recommended: FLUX.2 [klein]** - Fast iterations - Quick concept testing - Mood board generation - Exploring prompt variations **Production Marketing Assets:** **Recommended: FLUX.2 [pro]** - Consistent quality - Reasonable
aaif-gooseAGENTS.mdaaif-goose/goose/AGENTS.md
Until the migration is complete, changes to agent-loop behavior must be implemented and tested in both paths. When reviewing code, check whether a change to either path also applies
JetBrainsAGENTS.mdJetBrains/kotlin/analysis/AGENTS.md
ownership with `withValidityAssertion` - Mark experimental APIs with `@KaExperimentalApi`, implementation details with `@KaImplementationDetail` ## Working with Test Data When modifying test data files or running generated tests (`*Generated`) that compare output against
JetBrainsAGENTS.mdJetBrains/kotlin/compiler/AGENTS.md
mostly related to FIR (`compiler/fir/`), use `FIR: ` prefix in the commit subject line. - **Test-before-fix**: When fixing an issue and adding a test, commit the test data
router-for-meAGENTS.mdrouter-for-me/CLIProxyAPI/AGENTS.md
build -o cli-proxy-api ./cmd/server # Build go run ./cmd/server # Run dev server go test ./... # Run all tests go test -v -run TestName ./path/to/pkg # Run single test go build
ChromeDevToolsAGENTS.mdChromeDevTools/chrome-devtools-mcp/AGENTS.md
scripts from `package.json` to run commands. - Use `npm run build` to run tsc and test build. - Use `npm run test` to build and run tests, run all tests to verify
coreyhaines31AGENTS.mdcoreyhaines31/marketingskills/AGENTS.md
integration guides per tool │ └── REGISTRY.md # Tool index with capabilities ├── CONTRIBUTING.md ├── LICENSE └── README.md ``` ## Build / Lint / Test Commands **Skills** are content-only (no build step). Verify manually: - YAML frontmatter is valid - `name
AstralAGENTS.mdastral-sh/ruff/AGENTS.md
nitpicky. Report not only bugs and regressions, but also architectural and maintenance risks, weak test coverage, unclear code, unnecessary complexity, and meaningful style or consistency issues. Order findings by severity
ayghriAGENTS.mdayghri/i-have-adhd/AGENTS.md
plugin metadata. | | Documentation | `README.md`, `INSTALL.md`, `.github/readme/`, `.github/install/` | User-facing overview, installation, and translations. | | Verification | `tests/`, `scripts/` | Unit tests, compatibility checks, and evaluation tooling. | | Contribution workflow | `CONTRIBUTING.md`, `.github/pull_request_template.md` | Authorship, labels, safety
QuantumNousAGENTS.mdQuantumNous/new-api/AGENTS.md
appropriate when it represents reusable behavior, a required interface/framework callback, an exported API, a test fixture, or complex business logic that deserves direct tests. - If a single-use helper
QuantumNousAGENTS.mdQuantumNous/new-api/web/AGENTS.md
跨域与 Cookie 使用 `withCredentials` 并按后端要求处理 CSRF。 ### 3.14 测试 - 工具函数与纯逻辑优先单元测试（Vitest），测试文件 `*.test.ts`；组件用 React Testing Library 测交互与行为，避免测实现细节。 - 新增功能、修复缺陷或修改现有行为时，必须同步新增或更新测试；Bug 修复必须先编写能够稳定复现问题的失败用例，再实现修复并确认用例转为通过。 - 修改前端组件的布局、尺寸、滚动定位、焦点管理、键盘操作、选中状态
HKUDSAGENTS.mdHKUDS/nanobot/AGENTS.md
provider, executes tools, and manages session memory. ## Development Commands ```bash # Python: run single test / lint pytest tests/test_openai_api.py::test_function -v ruff check nanobot/ # Strict type checking (matches CI) uv sync
jeecgbootAGENTS.mdjeecgboot/JeecgBoot/jeecgboot-vue3/AGENTS.md
matching files under `src/`; run it only when broad formatting is intended. - No regular test script is configured in `package.json`; a Jest config exists, but there is no `test` script
sickn33AGENTS.mdsickn33/agentic-awesome-skills/AGENTS.md
docs/`; localized docs live in `docs_zh-CN/` and `docs/vietnamese/`. Maintenance scripts and tests are in `tools/scripts/` and `tools/scripts/tests/`. The hosted catalog app is in `apps/web-app/`. Registry outputs such
danny-avilaAGENTS.mddanny-avila/LibreChat/AGENTS.md
reject what is obsolete in a reply that says why. After each round: focused tests, `npx tsc --noEmit` in every workspace you changed, push, then request the next review naming
SentryAGENTS.mdgetsentry/sentry/AGENTS.md
Prefix the full relative path, or source the activate script: ```bash cd /path/to/sentry && .venv/bin/pytest tests/... cd /path/to/sentry && source .venv/bin/activate && pytest tests/... ``` For AI agents: always use `required_permissions
SentryAGENTS.mdgetsentry/sentry/src/AGENTS.md
scoped: `ProjectEndpoint` - Region silo: `RegionSiloEndpoint` 3. File locations: - Endpoint: `src/sentry/api/endpoints/{resource}.py` - URL: `src/sentry/api/urls.py` - Test: `tests/sentry/api/endpoints/test_{resource}.py` - Serializer: `src/sentry/api/serializers/models/{model}.py` 4. Document with drf-spectacular decorators
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.
