Summary

Addresses the need for controlled, collaborative AI programming assistance. It provides developers with an AI system that integrates intelligent decision-making, persistent memory, and semantic code search, transforming AI from a passive tool into an active collaborator. This enhances productivity and ensures AI actions align with project standards and user intent.

How It Works

Sanshu uses the Model Context Protocol (MCP) for AI-tool integration. Its core zhi (intelligence) tool acts as a mandatory interactive gateway, requiring user confirmation for critical AI decisions and enforcing principles like KISS/YAGNI/SOLID via the AURA-X-KYS protocol. The ji (memory) component offers global, persistent project memory, while sou (search) provides semantic code search with incremental indexing and smart waiting for up-to-date context.

Quick Start & Requirements

• Install: Download pre-compiled binaries from GitHub Releases for Windows, macOS, or Linux, and add the executable to your system's PATH. macOS users can also use Homebrew (brew install sanshu).
• Requirements: Rust 1.70+, Node.js 18+, pnpm 8+.
• Links: GitHub Releases

Highlighted Details

• Mandatory Interactive Gateway: The zhi tool ensures all critical AI decisions require explicit user approval, preventing autonomous actions.
• Persistent Project Memory: The ji component stores project-specific rules and context for consistent AI collaboration.
• Semantic Code Search: The sou tool offers natural language code querying with incremental indexing and smart waiting for real-time accuracy.
• Up-to-date Documentation: context7 provides access to the latest framework and library documentation, mitigating AI knowledge staleness.

Maintenance & Community

Utilizes GitHub Actions for automated CI/CD and releases, with contributions following the Conventional Commits specification. While specific community channels aren't detailed, a roadmap indicates ongoing development with planned features like a Prompt Enhancer and an ACE Public Pool.

Licensing & Compatibility

Released under the permissive MIT License, allowing commercial use, modification, and distribution without significant restrictions, making it compatible with closed-source projects.

Limitations & Caveats

The project is under active development, with several advanced features like a Prompt Enhancer and ACE Public Pool listed as planned. The CLI version is specifically optimized for certain environments (e.g., Windsurf), and while MCP compatibility is broad, full integration may depend on the client's MCP implementation.