CTX
Skill, agent, MCP, and harness recommendations for Claude Code/custom LLMs: 102,696-node LLM-wiki graph, 91,432 skills, 10,787 MCPs, 13 harnesses, and capped execution recommendations.
What is CTX?
CTX is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to skill, agent, mcp, and harness recommendations for claude code/custom llms: 102,696-node llm-wiki graph, 91,432 skills, 10,787 mcps, 13 harnesses, and capped execution recommendations.
Skill, agent, MCP, and harness recommendations for Claude Code/custom LLMs: 102,696-node LLM-wiki graph, 91,432 skills, 10,787 MCPs, 13 harnesses, and capped execution recommendations.
This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- Skill, agent, MCP, and harness recommendations for Claude Co
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx ctxConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use CTX
CTX (claude-ctx) is a knowledge-graph-backed recommendation engine for Claude Code and other LLM agents that surfaces the right skills, MCP servers, agent patterns, and harnesses for any coding task. It ships a 102,000-node LLM-wiki graph with 91,000+ indexed skills and 10,000+ MCP entries, so instead of searching GitHub manually you can ask CTX which tools best fit a given goal and get ranked, vetted recommendations. It also provides a local monitoring dashboard, pre-commit hooks, and health checks to keep an agent project well-organized over time.
Prerequisites
- Python 3.10+ with pip or uv installed
- Claude Code or another MCP-compatible AI client
- Optional: Docker (for some graph installation modes)
- Optional: An OpenAI or other LLM API key if using a custom model mode
Install the claude-ctx package
Install the core package via pip. Add the optional 'embeddings' extra to enable semantic search over the knowledge graph.
pip install claude-ctx
# or with semantic search support:
pip install "claude-ctx[embeddings]"Run the interactive setup wizard
Run ctx-init to walk through graph installation, hook setup, and model configuration. For a fast automated setup that downloads the full wiki graph and installs hooks, use the flags below.
# Interactive wizard
ctx-init
# Fast automated setup
ctx-init --graph --hooks --model-mode skip
# Full local graph installation
ctx-init --graph --graph-install-mode fullScan your repository for recommendations
Run ctx-scan-repo to analyze the current project and receive ranked recommendations for skills, MCPs, and agent patterns that fit the codebase.
ctx-scan-repo --repo .
ctx-scan-repo --repo . --recommendManage skills and MCPs
Use the add/quality/health commands to track which skills are registered in your project and check their quality scores.
ctx-skill-add my-skill
ctx-mcp-add my-mcp
ctx-skill-quality list
ctx-skill-health dashboardStart the local monitoring dashboard
Launch the CTX monitoring server to view a web dashboard of project health, skill usage, and recommendation history.
ctx-monitor serve
# Then open http://127.0.0.1:8765/ in your browserCTX Examples
Client configuration (Claude Desktop)
CTX is primarily a CLI tool and recommendation engine rather than a traditional stdio MCP server. After pip-installing claude-ctx, configure Claude Code to invoke ctx commands directly.
{
"mcpServers": {
"ctx": {
"command": "npx",
"args": ["ctx"]
}
}
}Prompts to try
Example requests you can make once CTX is available to your AI agent.
- "Scan my repo and recommend the best MCP servers for this project"
- "Which skills in the CTX graph are most relevant to building a REST API agent?"
- "Show me the skill health dashboard for this project"
- "Find harnesses that support multi-agent orchestration in the CTX graph"
- "Install the top-recommended harness for Claude Code projects"Troubleshooting CTX
ctx-init fails with 'graph download timeout' or network error
The full graph is large. Try the incremental mode first: ctx-init --graph --graph-install-mode incremental. If you are behind a proxy, set the HTTP_PROXY and HTTPS_PROXY environment variables before running the command.
ctx-scan-repo returns no recommendations
Ensure the knowledge graph is installed (check with ctx-skill-health dashboard). If the graph was not downloaded during ctx-init, re-run ctx-init --graph. Also verify you are pointing at a non-empty project directory with at least some source files.
Semantic search features are unavailable
Install the embeddings extra: pip install 'claude-ctx[embeddings]'. The base package ships keyword-only search; semantic ranking requires the additional sentence-transformers dependency included in the embeddings extra.
Frequently Asked Questions about CTX
What is CTX?
CTX is a Model Context Protocol (MCP) server that skill, agent, mcp, and harness recommendations for claude code/custom llms: 102,696-node llm-wiki graph, 91,432 skills, 10,787 mcps, 13 harnesses, and capped execution recommendations. It connects AI assistants to external tools and data sources through a standardized interface.
How do I install CTX?
Follow the installation instructions on the CTX GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with CTX?
CTX works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is CTX free to use?
Yes, CTX is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
CTX Alternatives — Similar Knowledge & Memory Servers
Looking for alternatives to CTX? Here are other popular knowledge & memory servers you can use with Claude, Cursor, and VS Code.
MemPalace
★ 52.6kA local AI memory system that stores all conversations verbatim and organizes them into navigable structures. It provides 19 MCP tools for AI assistants to search and retrieve past decisions, debugging sessions, and architecture debates automatically
Kratos
★ 25.7k🏛️ Memory System for AI Coding Tools - Never explain your codebase again. MCP server with perfect project isolation, 95.8% context accuracy, and the Four Pillars Framework.
Context Mode
★ 15.4kAn MCP server that preserves LLM context by intercepting large data outputs and returning only concise summaries or relevant sections. It enables efficient sandboxed code execution, file processing, and documentation indexing across multiple programm
Memu
★ 13.7kMemory for 24/7 proactive agents like OpenClaw.
MemOS
★ 9.3kMemOS (Memory Operating System) is a memory management operating system designed for AI applications. Its goal is: to enable your AI system to have long-term memory like a human, not only remembering what users have said but also actively invoking, u
Everos
★ 5.4kBuild, evaluate, and integrate long-term memory for self-evolving agents.
Browse More Knowledge & Memory MCP Servers
Explore all knowledge & memory servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up CTX in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use CTX?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.