Stop AI skills from silently breaking across platforms - Skills Compat Manager

AI agent skills (Claude Code, Cursor, OpenCode) are markdown files that look portable but aren't. A skill depending on bash or web_search silently fails on other platforms. Skills Compat Manager = CLI + MCP server: discovers skills from 6+ platforms, scans 4 dependency dims (pip/CLI/env/caps), AI-generates COMPAT.yaml (OpenAI/Claude/rules fallback), cross-platform matrix showing exactly what breaks where, MCP mode auto-injects compat info into every skill fetch. MIT, Python 3.11+, free.

Tempo - Align Work to Strategy — Turn Jira Data Into Strategic Clarity

Turn Jira Data Into Strategic Clarity

Hey Product Hunt! 👋 I'm Yuanheng, a PM and developer working with AI coding agents every day. I got frustrated building skills for Claude Code and watching them silently fail on Cursor or Claude.ai — there was simply no way to know what would break before it did. Skills Compat Manager was born from that frustration. It's a CLI tool + MCP server that: ✅ Scans your AI agent skills for cross-platform compatibility issues ✅ Auto-generates structured dependency declarations with AI ✅ Shows you exactly what will fail, where, and how to fix it ✅ Works as an MCP server so your agents get compatibility info automatically I built this for the growing community of AI agent power users who juggle multiple coding platforms. If you've ever had a skill work perfectly on one platform and mysteriously fail on another — this is for you. Open source (MIT), Python 3.11+, works with 6+ platforms. Feedback and contributions welcome! 🙌

Tempo - Align Work to Strategy — Turn Jira Data Into Strategic Clarity

Turn Jira Data Into Strategic Clarity