claude-code-organizer — Claude Code MCP server: what it does
A local dashboard that shows what your AI coding harness actually loads: memories, skills, agents, hooks, MCP servers and settings.
claude mcp add claude-code-organizer -- npx -y @mcpware/claude-code-organizerWhat claude-code-organizer does
CCO is an MCP server plus a local control panel for AI coding configs. You pick a harness in the sidebar and inspect what it reads — for Claude that means memories, skills, agents, hooks, slash commands, MCP servers and sessions; for Codex CLI, AGENTS instructions, custom agents, profiles, history, TOML config, plugins and skills; for OpenCode, layered JSON/JSONC config, commands, skill roots and themes; for DSH, profiles and layered entry files. Each harness goes through its own adapter, so precedence and mutation rules stay separate instead of being flattened into one model.
Markdown edits happen in place, inside the panel. Harness Doctor adds an effective context map and an explainable hygiene score, exact duplicates can be repaired reversibly, and skills move between roots as copies. There is also an MCP scanner: tool descriptions land in the model prompt, and a compromised server can bury instructions there that you would never notice. The Activation Scanner preview freezes a small local sensor model, reads its hidden activations and trains a probe before an untrusted capability runs.
Node.js 20 or newer is required. Telemetry is off unless you explicitly opt in: with it on, detailed event totals stay under ~/.cco/, and the outgoing monthly signal carries a rotating anonymous ID, the UTC month, the app version and a harness ID — no paths, names, prompts or file contents.
What it helps with
- see which configs a harness actually loads
- find and remove duplicate memories
- disable an MCP server for one project
- check MCP tool descriptions for hidden instructions
- move a skill to the right root as a copy
Example prompts for the agent
- Show me which memories and skills Claude is loading right now
- Find exact duplicates across my configs and list what can go
- Scan my MCP servers for suspicious tool descriptions
How to install claude-code-organizer
Platform CLI
claude mcp add claude-code-organizer -- npx -y @mcpware/claude-code-organizerPlatform CLI
codex mcp add claude-code-organizer -- npx -y @mcpware/claude-code-organizerConfig file
{
"mcpServers": {
"claude-code-organizer": {
"command": "npx",
"args": [
"-y",
"@mcpware/claude-code-organizer"
]
}
}
}Config file
{
"servers": {
"claude-code-organizer": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@mcpware/claude-code-organizer"
]
}
}
}Platform CLI
gemini mcp add claude-code-organizer npx -y @mcpware/claude-code-organizerStep by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about claude-code-organizer
How do I install claude-code-organizer in Claude Code?
Run claude mcp add claude-code-organizer -- npx -y @mcpware/claude-code-organizer. Codex CLI takes a similar step with its own command: codex mcp add claude-code-organizer -- npx -y @mcpware/claude-code-organizer.
What does claude-code-organizer do?
It puts the configs of several AI coding harnesses in one panel: what each tool reads, in-place Markdown editing, duplicate detection, and a check of MCP servers for poisoned tool descriptions.
What does CCO need to run?
Node.js 20 or newer. Telemetry stays off by default and needs explicit opt-in; once enabled, the monthly signal that leaves your machine has no paths or file contents.
