Purpose
This is a local-first terminal user interface tool that allows you to search your AI coding session history. It uses a hybrid approach of keyword and semantic retrieval to help you find past interactions across multiple AI assistants like Claude Code, OpenCode, and Codex.

Security Assessment
Overall risk: Low. The code scan reviewed 4 core files and found no dangerous patterns, hardcoded secrets, or requests for dangerous permissions. Because the tool indexes your local AI session history, it inherently accesses potentially sensitive data, such as your past coding queries, project files, and AI responses. However, it is designed to operate entirely local-first. The primary security consideration is that semantic search relies on an embedding model, which may require downloading a model or making network requests depending on your local configuration, so you should verify your specific setup to ensure data remains offline.

Quality Assessment
The project is licensed under the permissive MIT license and has clear, comprehensive documentation. It is actively maintained, with the most recent code push happening today. The main drawback is its low community visibility; with only 6 GitHub stars, it is a very new or niche project. Consequently, it has not yet been extensively peer-reviewed by the broader developer community.

Verdict
Safe to use, though you should remain aware that a low-star, early-stage project might undergo significant changes or lack extensive community support.