Skills
96,770 packaged capabilities and playbooks for agents. Install any of them with one command.
Testing & quality
Surface internal contradictions, conflicting requirements, and edge cases that expose hidden conflicts in a document, spec, plan, design, or set of requirements. Use when something feels off but you can't pinpoint why, before committing to a design, or when requirements have grown incrementally and may have drifted out of sync. TRIGGERS: 'consistency check', 'find contradictions', 'does this spec make sense', 'check for conflicts', 'something feels wrong here', reviewing requirements docs, technical specs, architecture plans, product briefs, or any document where internal coherence matters.
Use this skill to run the Shipshape Shipwright role: in-harbour code inspection. Discovers existing behaviour and policy violations from production code, writes @captain-tagged scenario skeletons and @planks annotations for Captain review. Run for fitting out, onboarding an existing codebase, and between releases.
Evaluate a skill against the legal skill design framework — thirteen design parameters (including trust surface, freshness, schema validation, and conflict detection), three legal failure modes, and a three-tier verdict (ready / some concerns / major concerns). Use before deciding whether to trust a community skill for installation, before deploying a first-party skill to the team, or when the user asks "should I trust this?" or "is this skill well designed?" Runs automatically as part of /legal-builder-hub:skill-installer.
Draft specs, analyze requirements, plan implementation, implement from specs, review against specs (generator≠evaluator), extract specs from existing code, run surgical bugfix specs, multi-session handoff/progress, and size greenfield/brownfield work. Use when the user wants testable requirements, living specs, spec-driven development, session handoff, bugfix with regression fences, Design-First vs Requirements-First, or /draft-spec /plan-impl /implement-spec /review /extract-spec. Do not use for C4/solution architecture packets (use solution-architecture), DevOps/CI-CD strategy design, README authoring, Agent Skill optimization, AFK/Ralph automation products, or trivial one-line fixes.
Use when a user-gate task has requiresUserSpecification=true OR the agent's "do I know HOW?" self-check returns no. Locks down verification mechanics through a short AskUserQuestion sequence and writes the answers back into the task's metadata. Does NOT run the verification itself.
Use when the user has a coding prompt — a feature, bug fix, or refactor — and wants it implemented end to end by a self-checking software loop, not a single pass. It refines the prompt into an executable plan (running the plan-loop internally), then executes the plan task by task: an Engineer subagent implements each PR-sized task editing only source, a separate QA subagent authors the tests that prove the task's acceptance criteria (or grades quality only when tests already cover them), runs the task's tests plus the full accumulated regression suite, and grades the code against a strict conciseness/readability/style-match rubric; the Engineer and QA loop on each task until the tests pass and the quality gate holds, then it commits and opens a pull request for the task — stacked PRs (one per task, each based on the prior) by default, or one combined PR — and moves to the next task. Author and critic have disjoint write scopes (Engineer owns source, QA owns tests) so neither can game the gate. Not for producing a plan without building it (that is plan-loop), and not for tuning a metric on an existing artifact under a fixed correctness bound (that is optimize-loop).
Everything to do with tests on HelpMeTest. Use when: writing tests for a new feature, generating tests for an existing feature, fixing a broken test, debugging a failing test, tests broke after a UI change, tests are out of date after a refactor. Triggers on: 'write tests', 'generate tests', 'test is failing', 'fix tests', 'tests broke', 'implement X', 'add feature', 'fix bug', 'why does this test fail', 'tests are out of date'. If it involves HelpMeTest tests in any way, this is the skill.
Test-driven development, including frontend slices verified against a Figma design source via Storybook stories and an agent-browser visual gate. Use when user wants to build features or fix bugs using TDD, mentions "red-green-refactor", or wants integration tests.
Use when writing, reviewing, or modifying unit/integration tests, adding test cases, assertions, or mocks, or deciding whether a piece of code is worth testing — especially when tempted to add tests for coverage, over-mock dependencies, or write formalistic tests that just restate the implementation.
End-to-end feature testing — browser QA, API verification, eval tests, or any combination. Covers browser interactions (via agent-browser CLI), Google Workspace operations (gws CLI), API calls, and LLM eval tests. Can also persist tests as reusable QA flows or eval files.
Automatically suggest tests for new functions and components. Use when new code is written, functions added, or user mentions testing. Creates test scaffolding with Jest, Vitest, Pytest patterns. Triggers on new functions, components, test requests, testing mentions.
Improve one existing shipped skill from concrete feedback or a failed run via bounded PDCA. Use when skill instructions or package behavior look wrong and the fix belongs in that one package; otherwise route to the owning workflow. Triggers: tune-skill.
Vitest testing framework. Covers unit tests, mocking, and coverage. Use for testing Vite-based and Node.js projects. USE WHEN: user mentions "vitest", "vite test", "unit test", asks about "vi.fn", "vi.mock", "test coverage", "mock functions", "test vite project" DO NOT USE FOR: E2E tests - use `playwright` instead; React component testing - combine with `testing-library`; Jest projects - use `jest` instead
Perform code review by comparing the current branch with the base branch, generate a complete audit plan and execute it item by item, covering all changed files, integrating EliteForge standards checks, security detection, business logic correctness verification, and reliability and solution soundness assessment, with support for interrupt recovery; results are saved to the `{OUTPUT_DIR}` directory.
Tester automation for Yunxiao (Projex): accept and start 【Test】 tasks and advance requirements from pending test → testing; record plans, test case execution, reports, and test deployment evidence from the verifiable test evidence checklist; after diagnosis and duplicate checking, create defects in one click; for both standalone defect creation and defect creation from test cases, the verifier is mandatorily set to the currently logged-in tester and read back, (this-cycle deliverables bind the development owner / non-this-cycle ones use a self-designated owner); pull the list of fixed | not-to-be-fixed pending verification; batch close fixed ones, reopen reproduced defects, and merge closed ones into the current iteration only after each Bug has independent retest-pass evidence; the normal completion path keeps the original evidence gate, and "complete testing" may also enter the audit release branch when the tester explicitly gives "manual confirmation passed = yes"; mark the current 【Test】 task as completed and advance the requirement to test complete, and mark it as an unofficial release candidate, only when all non-cancelled development tasks uniquely map to test tasks and sibling test tasks are complete. Use when the user says YunxiaoQA, test task, pull test tasks, create defect, reopen, batch close, merge into iteration, start testing, record test evidence, complete testing, manual confirmation passed, close the loop on test tasks, receive release backflow, verify release backflow, hand over release. Also call proactively when the user does not write the Skill name or standard phrase but natural language clearly expresses viewing their own test tasks, starting/continuing/completing formal testing, recording evidence, filing a Bug, retest-close, or reopening; first uniquely locate the 【Test】 task or Bug via the official CLI, then normalize to existing commands. With zero or multiple candidates, return read-only without guessing the number from the title. Do not trigger when it is only about running local code tests, developer self-testing, or fixing code with no Yunxiao test task context. Tester role only; does not create 【Development】/does not create iterations/does not fix already-fixed items on behalf of developers. For any write to Yunxiao, first Plan for confirmation, then apply in one go; do not align with yunxiao-requirement-lifecycle.
Independent third-person code review. Use after ALL build phases are complete for a holistic review of the entire feature, or anytime the user wants a quality gate. Reviews as an independent Senior Architect who did NOT write the code and now owns it, so it must meet world-class standards.
WCAG 2.1/2.2 compliance, WAI-ARIA implementation, screen reader optimization, keyboard navigation, and accessibility testing expert. Use PROACTIVELY for accessibility violations, ARIA errors, keyboard navigation issues, screen reader compatibility problems, or accessibility testing automation needs.
Use when making a repo agent-ready, verifying harness consistency, checking for documentation drift, bootstrapping harness infrastructure (AGENTS.md as index, docs/ structure, CI verification, enforcement mechanisms), or auditing repo agent-readiness maturity level.
Run a bounded second-model code review with Codex Sol or Claude Fable. Use one combined correctness, security, structural-simplicity, modularity, and reuse contract for one initial review plus up to two convergence retries when accepted findings change the source. Use when the user asks for agent review, a second opinion, security review, adversarial review, or deep code-quality review.
Clarifies the relationship between aggregates and transaction boundaries, and detects and corrects the anti-pattern of including multiple aggregates in a single transaction. Applies the principle that an aggregate is a strong consistency boundary, and that when a use case updates multiple aggregates it should use eventual consistency. Use when detecting transaction boundary issues during code review, use case design, or refactoring. Target languages: language-agnostic (Java, Kotlin, Scala, TypeScript, Go, Rust, Python, and all others). Triggers: activated by transaction boundary related requests such as "updating multiple aggregates in the same transaction", "the use case has @Transactional", "how to keep consistency between aggregates", "should we use the Saga pattern", "transaction boundary design", "one transaction one aggregate", "implementing eventual consistency", "transactions spanning aggregates".
Audit websites for SEO, performance, security, technical, content, and 15 other issue categories with 230+ rules using the squirrelscan CLI. Returns LLM-optimized reports with health scores, broken links, meta tag analysis, and actionable recommendations. Use to discover and assess website or webapp issues and health.
Use when the user has concrete failing cases in code or a guardrail/classifier/filter/prompt/API they own — a red-team failure catalogue OR a CI/CD test-failure report (failing pytest/JUnit tests) — and wants the target patched until those failures are closed without breaking what already works. It points straight at the failed cases (normalize any source with tools/ingest.py), fixes one root-cause class per iteration, and re-checks with tools/verify.py — oracle mode against a red-team oracle, or tests mode against the test suite — keeping a patch only if it closes a class while nothing that passed before regresses, else reverting; loops until every class is closed (dry) or the budget runs out, then opens a pull request with the patch set. The defensive fixer half of a find→fix setup. Not for discovering new failures (that is red-team), and not for editing the oracle, tests, or holdout that define ground truth.
Drive a real browser to QA a feature end-to-end as a user would. Loads the Playwright MCP playbook plus the failure modes to avoid. Use whenever you need to verify a UI feature works in a browser, capture PR screenshots, repro a customer bug visually, or do end-of-task dogfooding before declaring something "done". This is the QA stage of orchestrate mode.
Analyses code and produces a structured violation report against clean coding standards. Use when: auditing code before a refactoring task, reviewing a PR for clean coding compliance, or establishing a baseline before applying clean-code-refactor. Produces a violation report that clean-code-refactor and data-engineer Implement Mode can act on. Supports Python, JavaScript/TypeScript, C#, and Rust. Supports both general (Clean Code) and OB (BORO Quick Style Guide) convention sets via the `standard` parameter.
Generate and review tests following project testing standards. Use when: adding tests for a new or untested function/class, reviewing existing tests for quality compliance, or identifying untested paths in a module. Supports Python, JavaScript/TypeScript, C#, and Rust. Supports both general (Clean Code) and OB (BORO Quick Style Guide) convention sets via the `standard` parameter.
Prevent trivial helpers, wrapper layers, rename-only utilities, duplicate constants, and local reinvention. Enforce reuse of platform primitives, framework capabilities, shared utilities, and proven libraries. Use when adding or reviewing helper sprawl, object-vs-positional parameter decisions, arrow-vs-declaration choices, class-vs-functional design, god-file boundaries, schema-file boundaries, security/maintainability hardening, or UI structure so the agent keeps code disciplined and avoids technical debt.
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.