What problem does it solve? Finding, installing, and quality-checking AI agent skills across multiple CLI platforms (Claude Code, Codex, OpenCode) is fragmented and manual, with no central registry or consistent installation workflow. ## Core Features & Use Cases - Registry Search & Install: Semantically search 400k+ community skills and download them into the correct CLI skills directory (~/.claude/skills/, ~/.codex/skills/, ~/.opencode/skills/) with a single command. - Skill Creation: Generate new skills from GitHub repositories, natural-language prompts, or Office documents (PDF/PPT/Word) using an LLM backend. - Quality Evaluation & Analysis: Score skills on five dimensions (safety, completeness, executability, maintainability, cost-awareness) and build relationship graphs across local skills. - Use Case: You want a browser automation skill for Claude Code. Run /skillnet search browser automation, preview the top result in a temp directory, then install it to ~/.claude/skills/ and mirror it into your dotfiles repo. ## Quick Start Ask the assistant to search the SkillNet registry for a skill matching your need, for example: search SkillNet for a code review skill and install it for Claude.