How real projects brief Codex, Cursor and every other agent that reads AGENTS.md.
maziyarpanahiAGENTS.mdmaziyarpanahi/openmed/AGENTS.md
change should explain the local command, the CI gate, and the expected developer workflow. ## Coding Style & Architecture Rules Use Python 3.10+, 4-space indentation, and 88-character lines. Prefer `snake
VercelAGENTS.mdvercel/eve/AGENTS.md
AGENTS.md Guidance for coding agents (and humans) working in this repository. For setup, PR workflow, and release process, see [CONTRIBUTING.md](./CONTRIBUTING.md). ## About eve eve is a filesystem-first framework
weave-osAGENTS.mdweave-os/router/AGENTS.md
write. - **Non-tautological tests** — every test must assert behavior that breaks if prod code deleted. ## Layer model and import rules Three concentric layers. Imports flow inward only. ``` +-------------------------------------------------------------------+ | cmd/router/main.go (composition root
metalbear-coAGENTS.mdmetalbear-co/mirrord/AGENTS.md
tricky or non-obvious behavior clearly. Write comments for readers of the resulting code, not reviewers of the patch. Comments must stand alone after merge and describe the current behavior
magnitudedevAGENTS.mdmagnitudedev/magnitude/design/AGENTS.md
document merely to rationalize accidental implementation behavior. ## Maintenance and enforcement Changes to governed code should be reviewed for both implementation correctness and design conformance. Reviewers and agents must check that
magnitudedevAGENTS.mdmagnitudedev/magnitude/inference-v3/AGENTS.md
branch is the only pre-review signal. Bot feedback (CodeRabbit, pre-commit.ci) is triaged, not obeyed: verify each item against the code, reproduce it before adopting a suggested fix (its proposed
NVIDIAAGENTS.mdNVIDIA/cuml/cpp/agents.md
Code Review Guidelines - cuML C++/CUDA **Role**: Act as a principal engineer with 10+ years experience in GPU computing and high-performance numerical computing. Focus ONLY on CRITICAL and HIGH
NVIDIAAGENTS.mdNVIDIA/cuml/python/agents.md
Code Review Guidelines - cuML Python **Role**: Act as a principal engineer with 10+ years experience in machine learning systems and Python API design. Focus ONLY on CRITICAL and HIGH issues
ElasticAGENTS.mdelastic/elasticsearch-js/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
CloudflareAGENTS.mdcloudflare/cloudflare-docs/AGENTS.md
task, read `.flue/AGENTS.md` first.** That file covers the worker architecture, workflow routing, specialist agents (code review, conventions, style-guide), R2 state management, local dev scripts, and deployment. All bot-related
CloudflareAGENTS.mdcloudflare/cloudflare-docs/.flue/AGENTS.md
than resolving them. A failed judge publishes filtered specialist findings as unverified. ### Routing | Specialist | Reviews | | ------------- | ----------------------------------------------------------------------------------------------- | | `code` | Every added line of non-MDX files, and added lines inside code blocks
MicrosoftAGENTS.mdmicrosoft/SynapseML/AGENTS.md
narrowest relevant repository skill: [Scala changes](.github/skills/scala-code/SKILL.md), [local setup](.github/skills/synapseml-local-setup/SKILL.md), and [code review](.github/skills/code-review/SKILL.md). 4. For requests to make one or more issues or PRs "5/5", "200% ready", or merge
GitHubAGENTS.mdgithub/gh-aw/AGENTS.md
intent - Workflow create/update/debug/upgrade tasks → `.github/skills/agentic-workflows/SKILL.md` - Core engineering conventions, validation flow, and command playbooks → `.github/skills/developer/SKILL.md` - Code organization, file structure, WASM stubs, string patterns → `.github/skills/developer-code-organization/SKILL.md` - Security best practices, template injection, shell script
iflytekAGENTS.mdiflytek/skillhub/AGENTS.md
enum. Follow the design doc's intent for new code. **Key transitions:** - Normal user first upload → `PENDING_REVIEW` (no initial DRAFT) - SUPER_ADMIN first upload → `PUBLISHED` (direct publish) - Review approve
aipochAGENTS.mdaipoch/open-science/AGENTS.md
React adapter. Put a key in `common` only when its UI meaning and reviewed translation are the same in both processes; an identical English key is not enough. The guard
SaladDayAGENTS.mdSaladDay/cc-switch-cli/AGENTS.md
issues. - Validate every finding against the code. Make minimal fixes for confirmed issues within scope, run the relevant checks, then start another blind review with a new subagent under
tiannAGENTS.mdtiann/hapi/AGENTS.md
passing checks when code, dependencies, and environment are unchanged; commit/push/PR transitions alone do not require reruns. Run repository-wide checks when explicitly requested as well. - Review this task's changes
TabularisDBAGENTS.mdTabularisDB/tabularis/AGENTS.md
symbols, 25156 relationships, 300 execution flows). Use the GitNexus MCP tools to understand code, assess impact, and navigate safely. > Index stale? Run `node .gitnexus/run.cjs analyze` from the project root
clideyAGENTS.mdclidey/whodb/AGENTS.md
source of truth for Codex, Claude Code via import, opencode, Pi, and other compatible coding agents. - Use `.agents/README.md` as the index for deeper agent guidance, and read only the relevant
MistralAGENTS.mdmistralai/mistral-vibe/client-e2e/AGENTS.md
timing in those snapshot tests, it will always fail - The tests should be completely code agnostic - The tests should concsist of scenarios with just a list of events - user bytes
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.
