发现 Skills

Generate a Pundit authorization policy with spec file

End-to-end testing scenarios for Supabase — complete workflow tests from project creation to AI features, validation scripts, and comprehensive test suites. Use when testing Supabase integrations, validating AI workflows, running E2E tests, verifying production readiness, or when a user mentions Supabase testing, E2E tests, integration testing, pgvector testing, auth testing, or test automation.

Produces a concrete eval suite plan grounded in Microsoft's Eval Scenario Library and MS Learn agent evaluation guidance — scenario types, evaluation methods, quality signals, thresholds, and priority order — before any test cases are generated or evals are run.

This skill enables visual inspection of websites running locally or remotely to identify and fix design issues. Triggers on requests like "review website design", "check the UI", "fix the layout", "find design problems". Detects issues with responsive design, accessibility, visual consistency, and layout breakage, then performs fixes at the source code level.

Build multi-step, deterministic, step-gated Vectara agents. State machines via first_step + steps[] + next_steps + allowed_tools enforce phase transitions STRUCTURALLY at the platform level — not via session.metadata variables read by a single-step LLM prompt. Covers sub-agent delegation, structured-output gating, runtime-constrained generation, MCP/lambda tools, and cross-session approval workflows.

Headless CMS integration guidance — Sanity (native Vercel Marketplace), Contentful, DatoCMS, Storyblok, and Builder.io. Covers studio setup, content modeling, preview mode, revalidation webhooks, and Visual Editing. Use when building content-driven sites with a headless CMS on Vercel.

Skill for implementing features based on documentation, bridging technical specs (API, ERD, Stories) and code using a Test-Driven Approach.

Clone or update https://github.com/vibe-motion/procedural-fish and render procedural-fish animation to a video using the project's own render command. Use when the user asks to render 程序鱼/procedural fish, export a 程序鱼视频, or run procedural-fish Remotion rendering.

Skill imported from viblocks/viv-typed-agents

Check for glab CLI updates and view latest version information. Use when checking if glab is up to date or finding available updates. Triggers on update glab, check version, glab version, CLI update.

Contract review and legal analysis

Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.

Break the autocompact retry loop after several consecutive failures to avoid hammering shared APIs.

Run delist or blocklist enforcement after synchronization so removals update both metrics and downstream refresh signals.

Keep the universal skill compact yet discoverable by loading minimal front matter first and heavy content on demand.

Select the most useful memory files by comparing headers/descriptions via a JSON-schema-guided assistant check.

Resolve user-input attachment triggers before running later thread-safe attachment builders so dependent state is in place.

Delay surfacing max-output-token failures until the recovery loop completes, then emit sanitized tool_result errors.

基于字节跳动 Arco Design 的 Web UI 设计规范和最佳实践

Search and download Zenodo records, and manage authenticated Zenodo deposits through the API, including drafts, metadata, files, deletion, and publishing.

Complete pre-commit workflow - run quality checks (format, lint, type, test) and validate/create conventional commit messages

Mark Glazebrook (Dalhousie University, Canada) 视角. 跟腱病 PhD + 关节镜专家 + COA 前主席. 调用此 skill 时, 用 Glazebrook 框架做跟腱病变 / 关节镜 / 循证治疗等级决策.

Commits changes in the repository using Git VCS

Generate a long-form Chinese interview-prep cheat sheet on a specific ML/LLM topic — formulas with derivations, from-scratch PyTorch code, comparison tables, and 25 high-frequency interview questions (L1 must-know / L2 advanced / L3 elite lab). Cross-model codex review checks math, code, historical citations, and style discipline; then /render-html produces a single-file HTML with an academic-newspaper template. Output: docs/tutorials/<slug>_tutorial.{md,html,review.json}. Use when the user says "write an interview cheat sheet", "write an X tutorial", "help me prepare Y interview questions", "produce an X quick reference", or wants a 600–1000-line Chinese tutorial on a specific ML topic.