How real projects brief Claude Code: the commands, conventions and warnings they put in CLAUDE.md.
7onezCLAUDE.md7onez/cti-expert/CLAUDE.md
store ``` $INTEL_HOME/cases/ / # = intel_engine/cases/ / ``` The CLI (`scripts/backend/intel.py`), the pipeline (`intel_engine/tools/intel.py`) and the MCP server (`intel_engine/harness/tools.py`) **all** resolve to it. Confirm with `python3 scripts/backend/backend.py status`. **Never create
aitytechCLAUDE.mdaitytech/agentkits-marketing/CLAUDE.md
management & coordination - `copywriter` - Content creation & messaging - `docs-manager` - Marketing documentation & brand guidelines - `mcp-manager` - MCP server integrations & tool orchestration ### Reviewer Agents (Quality Assurance) - `brand-voice-guardian` - Brand consistency and voice
anombyte93CLAUDE.mdanombyte93/prd-taskmaster/CLAUDE.md
lockstep: - `script.py` — thin shim over `prd_taskmaster/cli.py:main` (~40 subcommands; `python3 script.py -h`). - `mcp-server/server.py` — FastMCP server exposing the same ops as MCP tools (`next_task`, `set_task_status
FradSerCLAUDE.mdFradSer/dotclaude/marketing/CLAUDE.md
introw, composio - **Composio** (integration layer): Adds MCP access to OAuth-heavy tools without native MCP servers (HubSpot, Salesforce, Meta Ads, LinkedIn Ads, Google Sheets, Slack, etc.). See `tools/integrations/composio.md` ### Registry Structure
christopherkaraniCLAUDE.mdchristopherkarani/Swarm/CLAUDE.md
Tool, @Parameter, │ │ # @Traceable, #Prompt, builders) │ ├── SwarmMembrane/ # Deprecated hollow re-export (remove in 0.7.0) │ ├── SwarmMCP/ # MCP server adapter product │ ├── SwarmCapabilityShowcase/ # Executable: deterministic showcase CLI │ ├── SwarmCapabilityShowcaseSupport/ # Library backing the showcase │ ├── SwarmDemo
minhnv0807CLAUDE.mdminhnv0807/ai-business-skills/CLAUDE.md
luong [Kiem tra truoc khi giao] ``` ## MCP Ads Integration Khi user da ket noi MCP server cho nen tang quang cao, cac skill co the pull data tu dong thay
AbilityaiCLAUDE.mdAbilityai/trinity/CLAUDE.md
agent runs as an isolated Docker container with standardized interfaces for credentials, tools, and MCP server integrations. **Local**: http://localhost **Backend API**: http://localhost:8000/docs --- ## Development Skills (`.claude` submodule) Skills
wuxiranCLAUDE.mdwuxiran/cc-pane/CLAUDE.md
的链路前先问：这个 id 是每次启动新生成的，还是稳定的？详见 docs/69。 - **stdio MCP 服务器不能给 `Stdio::null()` 的 stdin，它靠 stdin 活着**：`shared_mcp_service.rs::spawn_server_process` 对所有 bridgeMode 一律 null 掉三个流，stdio 类服务器于是启动后
rtk-aiCLAUDE.mdrtk-ai/icm/CLAUDE.md
Détermine l'importance d'une information async fn assess_importance(&self, content: &str) -> Result ; } ``` ## MCP Server Tools Le serveur MCP doit exposer ces tools pour Claude Code: ### icm_store ```json
HelmiCLAUDE.mdHelmi/claude-simone/CLAUDE.md
same time this repository itself uses Simone to organize its own work (currently the mcp-server based version). ## Activity Logging You have access to the `log_activity` tool
HelmiCLAUDE.mdHelmi/claude-simone/mcp-server/CLAUDE.md
CLAUDE Instructions - MCP Server Development ## Overview This is the MCP (Model Context Protocol) server implementation for Simone. This is a complete rewrite of the legacy directory-based command system. ## Tech
oliver-kriskaCLAUDE.mdoliver-kriska/claude-elixir-phoenix/CLAUDE.md
fields: `server`, `tool`; optional `input` with `${tool_input.field}` substitution. Caveat: SessionStart and Setup fire before MCP servers finish connecting, so for service detection prefer a direct probe (see `detect-tidewave.sh`); reserve `mcp
freddy-schuetzCLAUDE.mdfreddy-schuetz/n8n-claw/CLAUDE.md
workflow JSON files (source of truth) │ ├── n8n-claw-agent.json # Main agent workflow │ ├── mcp-builder.json # Builds new MCP Server workflows │ ├── mcp-client.json # Calls tools on MCP servers (sub-workflow) │ ├── mcp-library-manager.json # Installs/removes skills from template catalog
slopusCLAUDE.mdslopus/happy-cli/CLAUDE.md
Mobile app → Server → Claude SDK → Server → Mobile app 4. **Permission Handling**: - Claude requests permission → MCP server intercepts → Sends to mobile → Mobile responds → MCP approves/denies ## Key Design Decisions 1. **File-based
8beeeaaatCLAUDE.md8beeeaaat/touchdesigner-mcp/CLAUDE.md
test:integration` - Integration tests with TouchDesigner WebServer - `npm run test:unit` - Unit tests for MCP server components - `npm run test:python` - TouchDesigner-side Python tests against a stubbed `td` module
arlanrakhCLAUDE.mdarlanrakh/talk-to-girlfriend-ai/CLAUDE.md
Claude Code (claude.ai/code) when working with code in this repository. ## Project Overview Telegram MCP Server - a full-featured Telegram integration for Claude, Cursor, and MCP-compatible clients. Uses Telethon
fabio-rovaiCLAUDE.mdfabio-rovai/open-ontologies/CLAUDE.md
Claude adds them. The MCP tools are individual operations — Claude is the orchestrator. ## Tool Reference | Tool | When to use | | ---- | ----------- | | `onto_status` | To check if the server is running and healthy
nirholasCLAUDE.mdnirholas/XActions/CLAUDE.md
order: which lane to use (shell out to the `xactions` CLI, or load the MCP server), the CLI lane and its output contract, the MCP lane, a table from common
tzachbonCLAUDE.mdtzachbon/smart-ralph/CLAUDE.md
prompts - `/plugin-dev:hook-development` - Creating hooks (PreToolUse, PostToolUse, Stop, etc.) - `/plugin-dev:mcp-integration` - Integrating MCP servers into plugins - `/plugin-dev:plugin-settings` - Plugin configuration with .local.md files - `/plugin-dev:create-plugin` - Guided
butterbase-aiCLAUDE.mdbutterbase-ai/butterbase-skills/CLAUDE.md
right one per task: | Surface | What it's for | When to use it | |---|---|---| | **MCP tools** (this plugin) | Provisioning, schema, RLS, OAuth setup, function deploy, frontend deploy, integrations, billing config | Orchestration
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.
