stevesolun/ctx
A recommendation engine that analyzes user context and suggests a curated bundle of skills, agents, MCP servers, and harnesses for Claude Code and custom LLM projects from a 102K-node graph.
Not currently ranked — collecting fresh signals.
star history
The tool maintains a 102,928-node knowledge graph with 2.9M edges connecting skills, agents, MCP servers, and harnesses. It monitors what the user is building, walks the graph, and recommends the top-scored bundle of tools for the current task. For custom LLM setups, it provides a harness setup flow with dry-run, update, and uninstall controls. The graph includes 91,464 skill pages, 467 agents, 10,790 MCP servers, and 207 harnesses with semantic similarity, tags, and quality metrics.
Frequently asked
- What is stevesolun/ctx?
- A recommendation engine that analyzes user context and suggests a curated bundle of skills, agents, MCP servers, and harnesses for Claude Code and custom LLM projects from a 102K-node graph.
- Is ctx open source?
- Yes — stevesolun/ctx is open source, released under the MIT license.
- What language is ctx written in?
- stevesolun/ctx is primarily written in Python.
- How popular is ctx?
- stevesolun/ctx has 560 stars on GitHub.
- Where can I find ctx?
- stevesolun/ctx is on GitHub at https://github.com/stevesolun/ctx.