alvinunreal/lazyskills — reverse-engineered prompt

Reverse engineered prompt

Build me a really fast terminal app called LazySkills that acts like mission control for agent skills. I want one place where I can see every skill I have installed across project, global, universal, and agent specific locations, then switch between agents like Claude Code, Cursor, Codex, OpenCode, and Gemini CLI to see what is actually visible and usable.

It should make broken setups obvious, like missing SKILL.md files, bad frontmatter, broken symlinks, missing lock entries, and ghost skills. I want to open a skill, inspect its metadata and rendered content, preview the exact command before anything runs, and then safely update, reinstall, remove, or do bulk actions with confirmation prompts.

Please also include discovery so I can scan a local folder or a GitHub source for skills I have not installed yet. Make the keyboard driven TUI feel polished, with search, filters, detail views, key help, and an update flow from both the app and a simple CLI command. If you need details, look up current docs online.

Want more depth? Deep Reverse