How real projects brief Codex, Cursor and every other agent that reads AGENTS.md.
NVIDIAAGENTS.mdNVIDIA/NemoClaw/src/lib/messaging/AGENTS.md
when the manifest declares a channel policy preset. 7. Cover the behavior with manifest/compiler tests plus applier/onboard/channel CLI tests when host effects change. ## Where Changes Belong - New prompt, token, allowlist
claude-code-bestAGENTS.mdclaude-code-best/claude-code/AGENTS.md
files) bun run build # Build with Vite (alternative build pipeline) bun run build:vite # Test bun test # run all tests bun test src/utils/__tests__/hash.test.ts # run single file bun test --coverage # with
steipeteAGENTS.mdsteipete/CodexBar/AGENTS.md
XCTest coverage for usage parsing, status probes, icon patterns; mirror new logic with focused tests. - `Scripts`: build/package helpers (`package_app.sh`, `sign-and-notarize.sh`, `make_appcast.sh`, `build_icon.sh`, `compile_and_run.sh`). Release wrappers call `Scripts/mac-release`, which resolves
MicrosoftAGENTS.mdmicrosoft/onnxruntime/AGENTS.md
/code-review` skill in addition to relevant domain skills and matching path-scoped instructions. ## Build, Test, and Lint See the `/ort-build`, `/ort-test`, and `/ort-lint` skills for detailed instructions
Hugging FaceAGENTS.mdhuggingface/peft/.ai/AGENTS.md
submitting human is responsible for reviewing every changed line and running relevant tests. - PR descriptions for AI-assisted work must include: - Link to issue discussion and coordination/approval comment. - Which tests
ElasticAGENTS.mdelastic/kibana/AGENTS.md
each defined by a `kibana.jsonc`: core, packages, and plugin packages. Aside from tooling and testing, most code lives in these modules. - Packages are reusable units with explicit boundaries
GitHubAGENTS.mdgithub/docs/content/copilot/responsible-use/agents.md
ephemeral development environment where it can make changes to code, execute automated tests, and run linters. A firewall is enabled by default to prevent data exfiltration. * **Automated security scanning**: During
JetBrainsAGENTS.mdJetBrains/intellij-community/AGENTS.md
general guidance here. - **Product DSL** (`platform/build-scripts/product-dsl/`): follow its `AGENTS.md`. - **IJ Proxy MCP server** (`build/mcp-servers/ij-proxy/`): - Tests: run `bun run build` and `bun test`. - Bazel: do not run a Bazel build
JetBrainsAGENTS.mdJetBrains/intellij-community/java/AGENTS.md
code-placement` skill before you move code between them. - `intellij.java.tests` owns `testSrc` and `testData`. ## Tests ```bash ./tests.cmd --module intellij.java.tests --test
GoogleAGENTS.mdgoogle/filament/AGENTS.md
Operations, Verification, and Builds * [skills/verification_protocols/SKILL.md](skills/verification_protocols/SKILL.md): Mandatory verification pipeline (include formatting, desktop compilation, core test runs) before concluding a C++ task. * [skills/filament_build_clean/SKILL.md](skills/filament_build_clean/SKILL.md): Clean commands, debug desktop compilation, and release
MicrosoftAGENTS.mdmicrosoft/fluentui/AGENTS.md
dependency layers | [docs/architecture/layers.md](docs/architecture/layers.md) | ## Workflows | Topic | Location | | ------------------------------------ | ---------------------------------------------------------------- | | PR checklist, change files, commands | [docs/workflows/contributing.md](docs/workflows/contributing.md) | | Testing guide (unit, VRT, SSR, E2E) | [docs/workflows/testing.md](docs/workflows/testing.md) | | Team routing and label taxonomy | [docs/team-routing.md](docs/team-routing.md) | ## Quality
kortix-aiAGENTS.mdkortix-ai/suna/AGENTS.md
could solve it, what the tradeoffs are, and which option actually wins. **Then pressure-test it before you build:** - **Edge cases.** Which exist, which matter, which we can safely ignore
PydanticAGENTS.mdpydantic/pydantic-ai/AGENTS.md
known and loved for: - modern, idiomatic, concise Python - end-to-end type-safety and test coverage - thoughtful, tasteful, consistent API design - delightful developer experience - comprehensive well-written documentation In other
PydanticAGENTS.mdpydantic/pydantic-ai/docs/AGENTS.md
they keep the example representative. - Put example-level exclusions on the fence, such as `{test="skip" lint="skip"}`, rather than adding tooling suppressions to pedagogical code. Examples that are linted
PydanticAGENTS.mdpydantic/pydantic-ai/.github/workflows/AGENTS.md
itself must **not** have `needs: activation`. # A PR that edits `bots.yml` cannot test its own change `bots.yml` triggers on `pull_request_target`, and GitHub runs that trigger's workflow file
PydanticAGENTS.mdpydantic/pydantic-ai/pydantic_ai_slim/pydantic_ai/capabilities/AGENTS.md
CARRIED` in `durable_exec/temporal/_run_context.py` say which a field is; prove the behavior with a test in `tests/durable_exec
PydanticAGENTS.mdpydantic/pydantic-ai/pydantic_ai_slim/pydantic_ai/durable_exec/AGENTS.md
changing graph/tool/output/streaming/MCP behavior, check whether durable wrappers need matching updates and add workflow-level tests where external runtime behavior matters. - Sync stream (`run_stream_sync`) inherits via `self.run_stream`; missing
PydanticAGENTS.mdpydantic/pydantic-ai/pydantic_ai_slim/pydantic_ai/models/AGENTS.md
pydantic_ai/settings.py`, and give a new `Model` class a case in `tests/models/test_model_settings_support.py` — that test probes each class's outgoing request and fails when a list and the wire disagree. - Narrow
PydanticAGENTS.mdpydantic/pydantic-ai/pydantic_ai_slim/pydantic_ai/realtime/AGENTS.md
difference must be deliberate, documented in [docs/realtime/](../../../docs/realtime/), and covered by a parity test in `tests/realtime/`. ## Provider adapters - Parse provider frames through typed SDK models, never ad-hoc `dict` access
PydanticAGENTS.mdpydantic/pydantic-ai/pydantic_ai_slim/pydantic_ai/toolsets/AGENTS.md
feature-specific state on `Agent` when a toolset or capability can own it. - Test through public agent/toolset behavior where possible; snapshot message/tool-call history when behavior affects protocol shape
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.
