mcp-context-cache
v1.1.5
Published
MCP server for AI agent context caching
Maintainers
Readme
mcp-context-cache
🚀 MCP server for AI agent context caching
Efficiently loads, caches, and delivers project file context to AI agents using the Model Context Protocol (MCP).
Quick Start
# Install globally
npm install -g @tu-org/mcp-context-cache
# Generate configuration in your project
mcp-context-cache --initFeatures
- LRU Caching — Files are hashed (SHA-256) and cached. Unchanged files are never re-read.
- Security Pipeline — 8-step validation prevents loading sensitive files
- 3 MCP Tools — Flexible context loading: by file list, directory, or config file
- Smart Truncation — Respects size limits and reports omitted files
- Alphabetical Ordering — Deterministic output for stable AI responses
Documentation
- Full README — Installation, configuration, examples
- Agent Instructions — How AI agents use this server
- API Specification — Complete specification
- Contributing Guide — How to contribute
Development
pnpm install # Install dependencies
pnpm build # Build TypeScript
pnpm test # Run 43 tests
pnpm lint # Lint code
pnpm dev # Development modeLicense
MIT — See LICENSE
Built with TypeScript • MCP Protocol • LRU Cache • Zod
