liush2yuxjtu
@liush2yuxjtu
Public Skills
run
by liush2yuxjtu
Launch and drive the current project's real runtime surface so a human-visible or programmatic behavior can be observed. Use when asked to run the app, try a change, open the UI, exercise a CLI/TUI/API, or capture runtime evidence. Prefer an existing project run-* skill; otherwise infer the smallest safe launch path.
run-skill-generator
by liush2yuxjtu
Discover, verify, and persist a project-specific recipe that lets future agents build, launch, and drive the real application surface from a clean environment.
verify
by liush2yuxjtu
Verify ChatCut playable/mirror changes against the exact candidate using the repository's real Vercel browser acceptance harness, desktop/mobile interaction, hydration checks, and captured evidence.
html-artifacts
by liush2yuxjtu
Build a single self-contained, interactive HTML artifact and pop it open in the browser. Covers reports, slide decks, flowcharts/diagrams, design systems, component sheets, clickable prototypes, animation sandboxes, code-review / PR writeups, module maps, research explainers, status / incident reports, and throwaway editor UIs (triage board, feature-flag toggler, prompt tuner). Reach for this WHENEVER the user wants to SEE something instead of reading prose about it — phrases like "make an HTML page/artifact", "build me a deck / report / diagram / mockup / dashboard / prototype", "show me this as a webpage", "visualize this", "render this", "pop it open / 弹开看看", "做成一页 HTML", or any time the answer lands better as a spatial, interactive page than as markdown. Output is one self-contained .html saved to ~/.claude/artifacts/ and opened in the browser, styled with the Claude editorial design system and modeled on the bundled reference/*.html pattern library ("the unreasonable effectiveness of HTML"). NOT for shippable production UI / React components, and NOT a replacement for /talk-html when the user specifically wants a shareable gist recap link.