How real projects brief Codex, Cursor and every other agent that reads AGENTS.md.
bytedanceAGENTS.mdbytedance/deer-flow/scripts/AGENTS.md
later changes after it is part of the trusted base. Preapproved digests must be code-reviewed in that first change; after the corresponding file revision lands, promote the consumed digest
lobehubAGENTS.mdlobehub/lobehub/AGENTS.md
components, hooks, helpers, or types). Smaller, focused files are friendly to humans and agents. ### Code Review Before reviewing a PR / diff / branch change, read the **deep-review** skill. Ordinary review
stablyaiAGENTS.mdstablyai/orca/AGENTS.md
enforces # Writing Pull Requests Fill in [`.github/pull_request_template.md`](./.github/pull_request_template.md), written for a reviewer who has never seen this code: - No jargon — plain language, no internal shorthand. - The before and after
ElasticAGENTS.mdelastic/elasticsearch/AGENTS.md
default` for throwing an exception for unexpected values, or an assertion error if this code branch is unreachable. ## Types, Generics, and Suppressions - Prefer type-safe constructs; avoid raw types
OpenAIAGENTS.mdopenai/openai-cookbook/AGENTS.md
part workflows. Store secrets in environment variables such as `OPENAI_API_KEY`; never hard-code keys inside notebooks. ## Testing Guidelines Execute notebooks top-to-bottom after installing dependencies and clear
MicrosoftAGENTS.mdmicrosoft/ai-agents-for-beginners/AGENTS.md
Notebooks should execute without errors - README files should be clear and accurate - Follow existing code patterns in the repository - Maintain consistency with other lessons ## Additional Notes ### Common Gotchas 1. **Python
MicrosoftAGENTS.mdmicrosoft/ai-agents-for-beginners/translations/en/AGENTS.md
Notebooks should execute without errors - README files should be clear and accurate - Follow existing code patterns in the repository - Maintain consistency with other lessons ## Additional Notes ### Common Gotchas 1. **Python
ruvnetAGENTS.mdruvnet/ruflo/AGENTS.md
type`) | Type | Purpose | |------|---------| | `coordinator` | Orchestrates other agents | | `coder` | Writes code | | `tester` | Writes tests | | `reviewer` | Reviews code | | `architect` | Designs systems | | `researcher` | Analyzes requirements | | `security-architect` | Security design | | `performance-engineer` | Optimization | ### Task
career-ops-hqAGENTS.mdcareer-ops-hq/career-ops/AGENTS.md
career-ops project (parent/sibling repos, other codebases on the machine), knowledge from other Claude Code projects on the same machine, and cross-session inferences not written into an in-scope
career-ops-hqAGENTS.mdcareer-ops-hq/career-ops/web/AGENTS.md
finds the user's files, and it keeps your real pipeline out of it. ## Reviews and scope `web/` is maintained as first-party: functional and correctness fixes are very welcome
colbymchenryAGENTS.mdcolbymchenry/codegraph/AGENTS.md
summing per-turn assistant usage**, not `result.usage` (last-turn only in current Claude Code). See `docs/benchmarks/call-sequence-analysis.md`.) The mechanism that drives everything here: **an agent falls back to Read/Grep the instant
diegosouzapwAGENTS.mddiegosouzapw/OmniRoute/AGENTS.md
# OmniRoute agent guide > **Single source of truth.** This file holds ALL project rules, conventions
code-yeongyuAGENTS.mdcode-yeongyu/oh-my-openagent/.agents/AGENTS.md
only issue/PR triage with evidence reports | | `hyperplan/` | yes | Adversarial multi-agent planning | | `pre-publish-review/` | yes | 12-agent pre-publish release gate | | `get-unpublished-changes/` | NEW | Skill form
code-yeongyuAGENTS.mdcode-yeongyu/oh-my-openagent/AGENTS.md
REAL `senpi` binary through the drivers in `packages/omo-senpi/scripts/qa/`, which build their own isolated `SENPI_CODING_AGENT_DIR` and IGNORE a caller-provided one, so the real `~/.senpi/agent` is NEVER written
code-yeongyuAGENTS.mdcode-yeongyu/oh-my-openagent/.opencode/AGENTS.md
Adversarial multi-agent planning. 5 hostile category members cross-critique, lead synthesizes. | | `pre-publish-review/` | Nuclear-grade 12-agent pre-publish release gate. Runs `/get-unpublished-changes`, spawns ultrabrain agents per change
mem0aiAGENTS.mdmem0ai/mem0/AGENTS.md
template asks how the *code* was written; drafting the description with a model is fine. The disclosure is never held against you, it tells a reviewer where to look. Silence
mem0aiAGENTS.mdmem0ai/mem0/.github/AGENTS.md
enforces squash-only merges, linear history, no deletion, no force-push, and one approving review. Two status checks belong in its `required_status_checks` rule: | Context | Posted
gsd-buildAGENTS.mdgsd-build/get-shit-done/docs/AGENTS.md
agents** detailed in this section. Twelve additional shipped agents (pattern-mapper, debug-session-manager, code-reviewer, code-fixer, ai-researcher, domain-researcher, eval-planner, eval-auditor, framework-selector, intel-updater
mvanhornAGENTS.mdmvanhorn/last30days-skill/AGENTS.md
scripts/generate-plugin-index.py --check ``` 5. Open the PR, fill in their template, and wait for code-owner review. To roll out a later update in their catalog, bump the pinned
tw93AGENTS.mdtw93/Pake/AGENTS.md
this file, so edit `AGENTS.md` only. Local-only overrides (`CLAUDE.local.md`, `AGENTS.override.md`, `.claude/settings.local.json`) stay ignored. ## Code Conventions - No Chinese comments in any source (Rust / TypeScript / any file). Comments and identifiers
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.
