Skills
97,501 packaged capabilities and playbooks for agents. Install any of them with one command.
Testing & quality
Iteratively improve a Claude Code skill using the skill-reviewer agent until it meets quality standards. Use when improving a skill with multiple quality issues, iterating on a new skill until it meets standards, or automated fix-review cycles instead of manual editing.
WCAG 2.2 compliance, ARIA authoring patterns, screen reader and keyboard navigation, color contrast, and cognitive/neurodiversity accommodations. Use when auditing accessibility, fixing a11y violations, choosing ARIA roles, checking contrast ratios, or designing for ADHD, dyslexia, autism, or motor impairment.
Generate API test suites from route definitions across frameworks: auth, input validation, contract, k6 load testing, mocking, and OpenAPI-driven generation. Use when adding new APIs, auditing test coverage, or building regression suites.
Inspect and validate Blender assets technically and visually. Use for `.blend`, `.glb`, `.gltf`, `.fbx`, or `.obj` quality checks; topology and export review; evaluated triangle/material/hierarchy metrics; standardized multiview renders; fresh-import verification; or evidence-backed review of an agent-generated mesh.
Find duplicate, near-duplicate, and overlapping test cases across the test suite. Identifies exact matches, semantically similar tests, and tests that are subsets of other tests. Provides recommendations for which tests to keep, merge, or remove.
Find, classify, triage, re-implement and verify a source site's dynamic surface (APIs, search, forms, modals, media, tags, client-rendered and sheet-backed content) during a migration to a platform. Migration-bound — invoked by prepare-migration, replica, migrate and rollout, or standalone on an already-migrated site; never for redesign-only work.
Opens the running app in a browser and verifies that recent UI changes actually work. Use for any quick smoke test of recent work — “go”, “test in browser”, “check in browser”, “make sure it works”, “verify it works”, “did it work”, “works on mobile” — including when the user appends “...and make sure it works” to a UI request. For design critique, use web-design-guidelines
Diagnose runtime problems in Go programs: panics and stack traces, deadlocks, goroutine leaks, memory leaks, OOM kills, race reports, and debugging with delve and pprof. Use when: "debug this panic", "read this stack trace", "deadlock", "memory leak", "goroutine count growing", "OOM", "program hangs", "race detector output", "use delve". Not for: optimizing code that works (go-performance-review), writing new concurrent code (go-concurrency-review), failing test design (go-test-quality).
Apply the Holistic Testing Model evolved with PACTS (Proactive, Autonomous, Collaborative, Targeted, Structured) principles. Use when designing comprehensive test strategies for Classical, AI-assisted, Agent based, or Agentic Systems building quality into the team, or implementing whole-team quality practices.
Knip reference for detecting unused code in TypeScript and JavaScript. Detects unused files, exports, dependencies, types, enum members, and circular dependencies; supports auto-fix, monorepos (workspaces), catalogs (pnpm / Yarn / Bun catalogs), knip.json / knip.config.ts, plugins, ignore, reporters (SARIF), first cleanup, gradual adoption, and how Knip works
Prepare a Honcho change for a pull request to plastic-labs/honcho. Invoke before opening a PR, when drafting a PR body, when asked if a branch is PR-ready, or when filling the pull request template. Checks required tests and docs, then writes Description / Proofs / Fixes.
Generate ready-to-fill QA deliverables from markdown templates: test plans, manual test cases, automated Playwright specs, regression suites, and bug reports. Use when asked to draft or fill any of these artifacts from requirements or user stories. Keywords: test plan, test case, bug report, regression suite, playwright spec, test strategy, test design, template.
Use when reviewing/judging a published, consumer-facing API reference an integrating developer calls an API from — deciding if they can integrate from it. A gate, not authoring. Judges against a single-sourced 11-condition usability + contract-consistency bar: getting-started + auth reach a first call; every api-spec operation (incl. events/webhooks) is documented with purpose + typed params + a worked example + errors; errors first-class; rate-limits + pagination documented where applicable; versioning + deprecation/sunset stated; the reference is CONSISTENT WITH THE HANDED-IN api-spec (no drift, no fabricated endpoint — the load-bearing check); samples runnable; an amendment re-syncs to the changed contract. Emits exactly `VERDICT: approve|revise` plus actionable findings; approves a proportionally-sized reference, revises on a named gap. Not authoring it, not the end-user user-guide, not the developer-adoption guide, not the engineering api-spec (reviewing-api-spec).
Use when reviewing/judging a test plan / QA verification plan — deciding whether a tester can execute it and every behavior is covered. A gate, not authoring. Judges it against a single-sourced 10-condition bar: every handed-in upstream behavior (feature-spec/api-spec ops+errors, else PRD) has >=1 TRACEABLE case; functional levels fit; entry/exit testable (coverage floor + open-defect threshold); environments + test data specified; each case has preconditions + steps + observable result (a metric-threshold-on-a-dataset for ML) + traces-to; the catalog is RISK-WEIGHTED (coverage gap, padded/thin, OR combinatorial blow-up = a finding); each warranted non-functional type (perf/security/WCAG 2.2 AA/compat/i18n) carries a target; an amend is reviewed delta-scoped; nothing fabricated. Emits exactly `VERDICT: approve|revise` + actionable findings; approves a proportionally-sized plan, revises only on a named gap. Not for authoring it, the release runbook, design-review, or the test scripts.
Provides web vulnerability testing methodology distilled from 88,636 real-world cases from the WooYun vulnerability database (2010-2016). Use when performing penetration testing, security audits, code reviews for security flaws, or vulnerability research. Covers SQL injection, XSS, command execution, file upload, path traversal, unauthorized access, information disclosure, and business logic flaws.
Drive Chromium through a persistent JavaScript browser/page/locator API. Read scoped accessibility snapshots with managed refs and diffs; perform guarded, verified interactions with trusted input. Raw typed CDP remains available for deterministic operations and unsupported mechanics. Use to automate, inspect, or script single or multiple browser tabs. Requires browser-harness-js on PATH and a Chromium connection through the extension relay or remote debugging.
Audits code for design pattern opportunities and anti-patterns — identifies places where a specific GoF or architectural pattern would solve an observable problem, and flags misapplied patterns that add complexity without benefit. Generates fix prompts. Trigger phrases: "design patterns", "pattern check", "pattern review", "refactoring patterns", "pattern analysis".
Establishes protected, test-only RED evidence for observable behavior before Cook changes production. Use when the user invokes `/cut`, Mold hands off an approved behavior spec, Cook has no valid GateReceipt, or a user/Pasteurize reproduction needs adoption; do not use for closed non-behavior work or Press hardening.
Evaluates interfaces, components, screens, and flows against universal UX/UI principles (heuristics, UX laws, Gestalt, cognitive psychology, accessibility) and delivers concrete, prioritized improvements. Use whenever the user shares UI code, screenshots, components, or mockups and wants feedback — even if they don't use the words "critique" or "review". Also trigger when the user asks "what's wrong with this UI", "how can I improve this", "review my component", "does this look right", "give me feedback on this design", or shares any interface and asks for thoughts. Trigger for partial slices too (a single button, form, or card) — not only full screens.
Benchmark a third-party coding-agent harness against FrontierHarness Eval using Runta runtimes. Provisions a clean runtime bound to the GitHub repo under evaluation, installs the Harbor and Pier stacks needed for Terminal-Bench and DeepSWE tasks, freezes a golden checkpoint, runs tasks from identical fresh restores while saving trajectories as evidence, generates a comparison diagram, and builds a shareable report. Use when evaluating, benchmarking, scoring, or comparing a coding agent harness, or when the user mentions FrontierHarness, DeepSWE, Terminal-Bench, Harbor, Pier, golden checkpoints, or harness trajectories.
Invoke when the user needs to validate game balance, tune economy systems, analyze progression curves, evaluate difficulty scaling, or check reward pacing. Includes Monte Carlo simulation and statistical validation. Triggers on: "balance", "economy tuning", "progression curve", "difficulty scaling", "reward pacing", "Monte Carlo", "fairness check". Do NOT invoke for economy/monetization design (use game-economy-designer) or code review (use game-code-review). Part of the AlterLab GameForge collection.
Improve existing test suites into minimal, high-signal behavior-focused harnesses with approval-before-mutation, conformance checks, guarded validation, bounded repair, and auditable handoff statuses. Use when asked to improve, trim, rewrite, delete, review, or harden tests around public contracts, business logic, schemas, security behavior, failures, edge cases, readability, or maintainability.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.