How real projects brief Claude Code: the commands, conventions and warnings they put in CLAUDE.md.
rivet-devCLAUDE.mdrivet-dev/sandbox-agent/examples/CLAUDE.md
host directories into Docker example containers. - If an example needs tools, skills, or MCP servers, install them inside the container during setup. ## Testing Examples (ACP v2) Examples should be validated
fynnflueggeCLAUDE.mdfynnfluegge/agtx/CLAUDE.md
current project's config (enables its scripts and copy_files) ./target/release/agtx trust # Run the MCP server (global mode, or project-scoped with a path) ./target/release/agtx mcp-serve [path] # Serve
silversteinCLAUDE.mdsilverstein/minutes/CLAUDE.md
appropriate app bundle - TCC-sensitive desktop work (hotkeys, Screen Recording, Input Monitoring, Accessibility): `./scripts/install-dev-app.sh` - MCP server changes: `cd crates/mcp && npm run build` (compiles TS server + builds UI, then restart MCP
Open-Source-LegalCLAUDE.mdOpen-Source-Legal/OpenContracts/CLAUDE.md
commit hooks) pre-commit run --all-files ``` ### Frontend (React/TypeScript) ```bash cd frontend # Start development server (proxies to Django on :8000) yarn start # Run unit tests (Vitest) - watches by default yarn
JetBrainsCLAUDE.mdJetBrains/compose-hot-reload/CLAUDE.md
invalidated after reload | | `hot-reload-devtools` | Developer tools UI window | | `hot-reload-mcp` | MCP server for AI agent integration | | `hot-reload-test` | Test infrastructure and test plugins | | `tests/` | Functional integration
m0n0x41dCLAUDE.mdm0n0x41d/haft/CLAUDE.md
coding agent (Claude Code, Codex, OpenCode, Cursor), this CLI for manual access, and the MCP server (`haft serve`) for programmatic access. Standalone interactive agent (`haft agent`), TUI, and desktop wrappers
miscusi-peekCLAUDE.mdmiscusi-peek/cheatengine-mcp-bridge/CLAUDE.md
Table Lua Script`; in that case execute `dofile([[C:\path\to\cheatengine-mcp-bridge\MCP_Server\ce_mcp_bridge.lua]])` instead of pasting the full bridge. Success log: `[MCP v12.0.0] MCP Server Listening
vellum-aiCLAUDE.mdvellum-ai/vellum-assistant/skills/assistant-migration/references/claude.md
imported: - **Anthropic / Claude account login**: not migrated. The creator signs into Vellum independently. - **Connected MCP servers and integrations**: reconnect through Vellum's MCP setup and OAuth connect flows. Any tokens
sceneviewCLAUDE.mdsceneview/sceneview/CLAUDE.md
demo app per platform — `android-demo` is the Play Store app | | `mcp/` | `sceneview-mcp` server + packages | | `agents/` | Installable SceneView skills — a product surface, shipped to users | Full API reference: [`llms.txt
sipyourdrink-ltdCLAUDE.mdsipyourdrink-ltd/bernstein/CLAUDE.md
system (pluggy) - `src/bernstein/plugins/` | | `tui/` | Textual TUI - `src/bernstein/tui/` | | `github_app/` | GitHub App integration - `src/bernstein/github_app/` | | `mcp/` | MCP server - `src/bernstein/mcp/` | | `benchmark/` | SWE-bench - `src/bernstein/benchmark/` | ## Directory context files Per-directory context files for agents
caliber-ai-orgCLAUDE.mdcaliber-ai-org/ai-setup/CLAUDE.md
opencode/index.ts` · `github-copilot/index.ts` · `refresh.ts` · `staging.ts` · `backup.ts` · `manifest.ts` · `pre-commit-block.ts` **Scanner** (`src/scanner/`): `index.ts` — detects local MCP servers, rules, skills and enabled plugins across platforms **Sync** (`src/sync/`): `types.ts` · `adapters.ts` · `reconcile.ts` · `state.ts` · `frontmatter.ts
modu-aiCLAUDE.mdmodu-ai/moai-adk/CLAUDE.md
interrupted agent work using agentId (e.g., "Resume agent abc123 and continue the analysis"). --- ## 12. MCP Servers & Deep Analysis Modes - **UltraThink** (`ultrathink` keyword) / **Adaptive Thinking** (Opus 4.7+, incl. Opus 5/4.8): sets
MicrosoftCLAUDE.mdmicrosoft/skills-for-fabric/CLAUDE.md
Lakehouse data - XMLA/DAX for Semantic Models - KQL for Real-Time Intelligence ### Consumers - Use **MCP servers** for natural language queries - Limited to: Semantic Models, Warehouses, Lakehouse SQL Endpoints - No ODBC/JDBC setup
Chorus-AIDLCCLAUDE.mdChorus-AIDLC/Chorus/CLAUDE.md
auth, prisma, api-response, uuid-resolver) ├── services/ # Business logic layer (all UUID-based) ├── mcp/ # MCP Server factory + permission-gated tool modules │ ├── server.ts # Creates per-auth MCP server instance │ └── tools/ # public.ts
affaan-mCLAUDE.mdaffaan-m/agentshield/CLAUDE.md
AgentShield Security auditor for AI agent configurations (Claude Code, MCP servers, hooks, agents). ## Build & Test ```bash npm run build # tsc + tsup → dist/ npm test # vitest (912 tests
cookjohnCLAUDE.mdcookjohn/zotero-mcp/zotero-mcp-plugin/CLAUDE.md
Zotero MCP Plugin ## Project Overview A Zotero plugin that provides MCP (Model Context Protocol) server functionality, enabling AI assistants to interact with Zotero's library data. ## Tech Stack - TypeScript - Zotero
AgriciDanielCLAUDE.mdAgriciDaniel/banana-claude/CLAUDE.md
plugin-dir . plugin details banana-claude@inline` to confirm the expected skill, agents, and MCP server when Claude Code is present. Live acceptance is a separate manual gate using
hoangsonwwCLAUDE.mdhoangsonww/Claude-Code-Agent-Monitor/CLAUDE.md
printed command, surfaced in the UI and by `ccam update-check`.) - `mcp/`: local MCP server exposing dashboard operations as tools. ## Non-negotiable engineering rules - Preserve existing behavior unless explicitly asked
bytechefhqCLAUDE.mdbytechefhq/bytechef/CLAUDE.md
both `com.fasterxml.jackson` and `tools.jackson` packages currently appear in deps), SpringDoc OpenAPI 3.0.3 ### Main Server Module Structure #### Core Modules (`server/libs/`) - **`atlas/`** - Workflow engine core - `atlas-coordinator/` - Orchestrates workflow execution - `atlas-execution
jarrodwattsCLAUDE.mdjarrodwatts/claude-delegator/CLAUDE.md
uninstall to test removal flow /claude-delegator:uninstall ``` No build step, no dependencies. Uses native MCP servers from Codex and Gemini CLIs. ## Architecture ### Orchestration Flow Claude acts as orchestrator—delegates
A file Claude Code reads at the start of every session. It holds the commands, conventions and warnings the agent needs for this project.
At the repository root. Claude Code also reads CLAUDE.md files in subdirectories when it works there.
Build and test commands, the project's layout, conventions that aren't obvious from the code, and mistakes to avoid. Short files tend to work better than long ones.
Claude Code reads CLAUDE.md; most other agents read AGENTS.md. Many projects keep one and point the other at it.
