全部能力

Full-stack diagnostic for agent and LLM applications. Audits the 12-layer agent stack for wrapper regression, memory pollution, tool discipline failures,…

Use when the user asks about finding the best, top, or recommended model for a task, wants to know what AI model to use, or wants to compare models by benchmark scores. Triggers on: "best model for X", "what model should I use for", "top models for [task]", "which model runs on my laptop/machine/device", "recommend a model for", "what LLM should I use for", "compare models for", "what's state of the art for", or any question about choosing an AI model for a specific use case. Always use this skill when the user wants model recommendations or comparisons, even if they don't explicitly mention HuggingFace or benchmarks.

Experimental NotebookLM harness for listing notebooks, managing sources, asking questions, generating artifacts, and downloading outputs through an installed…

Advanced AI agent benchmark scenarios that push Vercel's cutting-edge platform features — Workflow DevKit, AI Gateway, MCP, Chat SDK, Queues, Flags, Sandbox,…

Latest AI models reference - Claude, OpenAI, Gemini, Eleven Labs, Replicate

Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express…

Create and launch benchmark test projects to exercise vercel-plugin skill injection across realistic scenarios. Sets up isolated directories, installs the…

axiom-ios-games — an installable skill for AI agents, published by charleswiltgen/axiom.

End-to-end benchmark suite for vercel-plugin. Runs realistic projects through skill injection, launches dev servers, verifies everything works, analyzes…

Scaffolds a production-ready TypeScript CLI project with ESM, tsdown, vitest, oxlint, oxfmt, changesets, GitHub Actions, and an agent skill definition. Use…

Patterns for building AI agents that learn from their own execution, detect failure modes, and improve autonomously. Covers feedback loops, performance regression detection, memory curation, skill extraction, and meta-learning architectures. Use when building agents that need to get better over time, managing auto-memory, or designing self-correcting systems.

Run vercel-plugin eval scenarios in Vercel Sandboxes instead of local WezTerm panels. Provisions ephemeral microVMs with Claude Code + plugin pre-installed,…

Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express…

Integrate Lighthouse CI accessibility audits. Detects configuration, parses results, maps findings to severity model, and tracks score regressions.

Integrate GitHub Accessibility Scanner. Detects configuration, parses scanner issues, correlates with local scans, and tracks Copilot-assigned fixes.

Mechanically replace static dependency call sites with wrapper or built-in abstraction calls across a bounded scope (file, project, or namespace). Performs codemod-style bulk replacement of DateTime.UtcNow to TimeProvider.GetUtcNow(), File.ReadAllText to IFileSystem, and similar transformations. Adds constructor injection parameters and updates DI registration. USE FOR: replace DateTime.UtcNow with TimeProvider, replace DateTime.Now with TimeProvider, migrate static calls to wrapper, bulk replace File.* with IFileSystem, codemod static to injectable, add constructor injection for time provider, mechanical migration of statics, refactor DateTime to TimeProvider, swap static for injected dependency, convert static calls to use abstraction, replace statics in a class, migrate one file to TimeProvider, scoped migration, update call sites. DO NOT USE FOR: detecting statics (use detect-static-dependencies), generating wrappers (use generate-testability-wrappers), migrating between test frameworks.

This skill should be used when the user asks to "run npm test after 30 minutes", "git commit after 1 hour", "wait 2h then deploy", "sleep 45m and run build",…

Fixes GitHub issues using parallel analysis agents for root cause investigation, code exploration, and regression detection. Reads issue context from gh CLI,…

Implement Antithesis workloads by turning the property catalog into SDK assertions and test commands, then refine coverage after triage.

Specification-driven feature development with auto-sized depth. Produces spec.md, design.md, and tasks.md artifacts with requirements traceability, plus verify and audit phases tied to acceptance criteria. Use when planning a feature, breaking a change into tasks or stories, implementing a named story or task, verifying implementation against acceptance criteria, auditing goals before closing, or turning a PRD into engineering artifacts. Triggers: "plan this feature", "spec this feature", "turn this PRD into a spec", "break this into tasks/stories", "create technical design", "implement story S-1", "implement task T-1", "verify implementation", "check acceptance criteria", "audit feature", "audit this spec", "validate goals", "UAT", "manual testing", "discuss this feature", "show feature status", "quick fix", "quick task", "quick mode", "small change", and known one-line fixes where the user names file and line. Not for diagnosing unknown bugs, authoring standalone PRD/RFC/ADR/TDD documents, PR/commit mechanics, or PM backlog tracking.