How real projects brief Codex, Cursor and every other agent that reads AGENTS.md.
rpamisAGENTS.mdrpamis/comet/AGENTS.md
读取从仓库根到该文件目录沿途的 `AGENTS.md`；路径更深的文件补充或覆盖上层规则。`.claude/rules/*.md` 是 Claude Code 的对应路径规则，`.codex/rules/*.rules` 只用于 Codex 命令执行策略，不承载开发行为说明。 规则文件的数字前缀只用于分组和稳定排序，不表示优先级。发生冲突时按平台自身的作用域规则处理，不根据编号判断覆盖关系。 ## 问题证据与验证结论 Issue、Review 意见、Project Knowledge、Memory 和历史记录只提供调查线索。可能变化的事实必须在当前分支重新核对；行为缺陷应先在最小临时项目、对应发布包或真实
stravuAGENTS.mdstravu/crystal/AGENTS.md
patches. - Update docs alongside code; do not alter build targets without discussion. - Use repository scripts (pnpm) and keep formatting consistent with existing files. - Always review the root `CLAUDE.md` before beginning
Chachamaru127AGENTS.mdChachamaru127/claude-code-harness/codex/AGENTS.md
harness 開発ガイド このファイルは Codex CLI がこのリポジトリで作業する際の指針です。 ## プロジェクト概要 **Harness** は、Codex CLI を「Plan → Work → Review」の型で運用するためのガイドです。 **特殊な点**: このプロジェクトは「ハーネス自身を使ってハーネスを改善する」自己参照的な構成です。 ## Codex CLI の前提 - Codex は `${CODEX_HOME:-~/.codex}/skills/ /SKILL.md
Chachamaru127AGENTS.mdChachamaru127/claude-code-harness/.grok/AGENTS.md
smoke target only): - Worker → Grok subagent / spawn with worker role routing - Reviewer → read-only review skill / subagent - Advisor → advisor-request.v1 only - Multitask / background agents → optional parallel
Chachamaru127AGENTS.mdChachamaru127/claude-code-harness/opencode/AGENTS.md
/work" | | breezing | Full parallel run with Agent Teams | "run with team", "breezing" | | harness-review | Code review, quality checks | "review", "security", "performance" | | harness-plan | Planning and task shaping into Plans.md | "plan
Forward-FutureAGENTS.mdForward-Future/loopy/AGENTS.md
source of truth for public loops. The current Git tree holds application code and the content-free site shell. Do not commit published loop records, bootstrap data, generated loop pages
oceanbaseAGENTS.mdoceanbase/seekdb/AGENTS.md
once` in new C++ header files instead of macro-based `#ifndef`/`#define` include guards. ## Code Review - For pull request or diff review tasks, read and follow `.agents/skills/code-review/SKILL.md`. ## Test Placement - Prefer
MicrosoftAGENTS.mdmicrosoft/Generative-AI-for-beginners-dotnet/AGENTS.md
Validation is done through: - Build verification in CI - Manual testing of sample applications - Documentation review ## Code Style and Conventions ### Formatting ```bash # Format code before committing (REQUIRED) dotnet format # Verify formatting
jakubkrehelAGENTS.mdjakubkrehel/make-interfaces-feel-better/AGENTS.md
guidance to Codex (Codex.ai/code) when working with code in this repository. ## What this repository is A standalone agent skill for reviewing and improving interface polish, published for installation
elie222AGENTS.mdelie222/rakazo/AGENTS.md
secrets, `.env` files, private URLs, personal/customer data, or real production data; use fake placeholders. Review `git status` and the staged diff before committing, and never force-add ignored files
VercelAGENTS.mdvercel-labs/opensrc/AGENTS.md
AGENTS.md Instructions for AI coding agents working with this codebase. ## Monorepo Structure This is a **Turborepo** monorepo with pnpm workspaces: - `packages/opensrc/` — Rust CLI distributed via npm - `apps/docs/` — Next.js documentation site
rlaopeAGENTS.mdrlaope/oh-my-hermes/AGENTS.md
multiple focused commits inside the same goal PR when useful. Planning docs, tests, implementation, code-review fixes, CI fixes, and small follow-up docs belong in the same PR when
NVIDIAAGENTS.mdNVIDIA/NeMo-Retriever/AGENTS.md
open that PR in the same task and link it from the code PR. - If the host cannot run subagents, read `docs/AGENTS.md` in the primary task, complete the documentation work
intellectronicaAGENTS.mdintellectronica/ruler/AGENTS.md
Unit tests - Integration tests - End-to-end tests ### Code Formatting - Always ensure that the code is formatted correctly by running prettier before committing changes. ### Branches and Pull Requests - Unless specifically
intellectronicaAGENTS.mdintellectronica/ruler/.ruler/AGENTS.md
Unit tests - Integration tests - End-to-end tests ### Code Formatting - Always ensure that the code is formatted correctly by running prettier before committing changes. ### Branches and Pull Requests - Unless specifically
GitHubAGENTS.mdgithub/copilot-cli-for-beginners/AGENTS.md
debugging exercises (Ch 03) | | `samples/agents/` | Agent template examples (python-reviewer, pytest-helper, hello-world) | | `samples/skills/` | Skill template examples (code-checklist, pytest-gen, commit-message, hello-world) | | `samples/mcp-configs/` | MCP server configuration
makecindyAGENTS.mdmakecindy/cindy/AGENTS.md
的重试／超时／断链恢复逻辑前，必须先读 `docs/dev-rules/remote-and-mobile-adaptation.md`；其中恢复路径改动必须回答该文件的 「故障半径三问」。 - 在 Cindy 内嵌 worktree 会话里工作、准备提交或直推、或做 code review 前，必须先读 `docs/dev-rules/development-workflow.md`。 ## 通用工作流程 1. 先确认用户目标、仓库边界、当前分支、worktree 和工作区状态。 2. 尊重开发者或宿主已经提供的 Git 工作流
Vexa-aiAGENTS.mdVexa-ai/vexa/AGENTS.md
history.** No "this used to be X", no bug archaeology in comments — write code as if it had always been this way. - **Report facts, then your reading — labelled as yours
openlitAGENTS.mdopenlit/openlit/AGENTS.md
seats, and entitlements are organisation-scoped in enterprise. ## UI and Security - Do not hard-code user-facing strings. Add CE strings to `src/client/src/constants/messages/en.ts`; enterprise-only strings belong in the private
MicrosoftAGENTS.mdmicrosoft/Ontology-Playground/AGENTS.md
conventional commits** format: - `feat:` - new features - `fix:` - bug fixes - `docs:` - documentation changes - `refactor:` - code refactoring - `test:` - adding or updating tests - `chore:` - maintenance tasks - **Write descriptive commit messages** with a summary
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.
