AGENTS.md examples from real projects

basicmachines-coReads credentialsAGENTS.md

basic-memory

basicmachines-co/basic-memory/AGENTS.md

Import functionality for Claude, ChatGPT, and other sources - `/markdown` - Markdown parsing and processing - `/mcp` - MCP server + `container.py` composition root + `clients/` typed API clients - `/models` - SQLAlchemy ORM models - `/picoschema` - Picoschema parsing

nubjsAGENTS.md

nub / wiki

nubjs/nub/wiki/agents.md

Claude: `.claude/settings.json` + `.claude/hooks/` (Claude Code), `.codex/config.toml` + `.codex/hooks.json` (Codex), and `.mcp.json` (the shared `chrome-devtools` MCP server). The two agents read different config files, so the `gh-comment-guard` hook script

archestra-aiReads credentialsAGENTS.md

archestra / platform

archestra-ai/archestra/platform/AGENTS.md

gives the Next dev server back (about 3.5 GB once pages are open). MCP server installs, nested Agent Runtime, the code runtime, Vault, Keycloak specs and the `api-k8s`, `hibernation

modelscopeReads credentialsAGENTS.md

ms-agent / webui

modelscope/ms-agent/webui/AGENTS.md

another terminal: cd frontend pnpm install --frozen-lockfile pnpm dev # frontend development server, :5173 ``` For the complete app, build with `pnpm build` in `frontend/`, then run `uv run webui

nyldnReads credentialsAGENTS.md

claude-octopus

nyldn/claude-octopus/AGENTS.md

Alias | |-----------------|---------------| | Read | read | | Write, Edit | edit | | Bash | execute | | Grep, Glob | search | ## MCP Integration The MCP server (`mcp-server/`) exposes Claude Octopus workflows as MCP tools. For MCP-aware coding agents

No other project uses it yet

ZenPrivacyAGENTS.md

zen-desktop

ZenPrivacy/zen-desktop/AGENTS.md

follow `.agents/skills/writing-commit-messages/SKILL.md` - The app's UI can be driven via a browser (e.g. using MCP) for debugging and testing features, including starting and stopping the proxy, and modifying filter lists

camundaDeletes or force-pushesAGENTS.md

camunda

camunda/camunda/AGENTS.md

engine | | `clients/` | Client libraries (Java, Spring Boot starters) | | `gateways/` | Gateway implementations (HTTP mapping, MCP) | | `security/` | Security core, protocol, validation | | `qa/` | Cross-component acceptance tests | | `testing/` | Process testing libraries | | `authentication/` | OIDC

UsefulSoftwareCoAGENTS.md

executor

UsefulSoftwareCo/executor/AGENTS.md

Executor policy Stable contracts live here. Current setup and server mechanics are in [RUNNING.md](RUNNING.md); e2e invariants are in [e2e/AGENTS.md](e2e/AGENTS.md). Run `bun run bootstrap` in every fresh rift. ## Verification

UsefulSoftwareCoAGENTS.md

executor / e2e

UsefulSoftwareCo/executor/e2e/AGENTS.md

test failures must not leak state. - Assert through the real typed API, browser, MCP, CLI, emulator ledger, or exported telemetry surface. Do not reach into implementation internals to make