Skills
97,222 packaged capabilities and playbooks for agents. Install any of them with one command.
Testing & quality
WordPress admin UI review and development guidance. Use when reviewing admin pages, settings screens, menu pages, notices, list tables, metabox-like interfaces, screen options, capability checks, admin scripts or styles, or when user mentions "admin UI", "settings page", "wp-admin", "admin screen", "options page", "list table", "admin menu", "admin notice", "screen options", or "dashboard page". Detects capability issues, Settings API misuse, admin-only performance problems, and UX mistakes in WordPress admin interfaces.
Write docs/<feature>/ACCEPTANCE.md — the ordered checklist a person works through by hand to deploy, test and accept finished work, because a machine could not do it. Use at the end of every implementing skill (execute-plan, execute-roadmap, build-feature), or when the user invokes /write-acceptance <feature> or asks for a test plan, an acceptance checklist, or "tell me how to verify this". Covers only what could not be automated — consoles and accounts, physical devices, GUI surfaces, two-machine and multi-user cases — with copy-pasteable commands, the exact observable that counts as evidence, and what you expect back. The reader may be the user or a tester who has never seen the feature. For handing unfinished work to the next agent, use write-handoff instead.
REDIRECT ONLY. Canonical xy-development lives in ariestools/ariestools-skills. Install that pack and read its xy-development skill for TypeScript conventions, Git workflow, testing principles, and the definition of done. This xyo-skills copy is a temporary redirect stub — do not use it as documentation.
REDIRECT ONLY. Canonical xy-toolchain lives in ariestools/ariestools-skills. Install that pack and read its xy-toolchain skill for @ariestools/toolchain, the xy CLI, ESLint flat configs, TypeScript configs, Vitest, deplint, repository policy, skills, and work tracking. This xyo-skills copy is a temporary redirect stub — do not use it as documentation.
AISBench Benchmark - AI model evaluation tool for Ascend NPU. Supports accuracy evaluation (service/local models on text, multimodal datasets), performance evaluation (latency, throughput, stress testing, steady-state, real traffic simulation), vLLM/Triton inference services, 15+ benchmarks (MMLU, GSM8K, MMMU, docvqa, ocrbench_v2, etc.), multi-turn dialogue, Function Call (BFCL), and custom datasets.
Manage Apidog project resources through Apidog CLI. Trigger scenarios - running API automation tests/test suites, querying/creating/updating/deleting endpoints, environments, schemas, mocks, branches and other project resources, importing/exporting API documentation, viewing test reports, managing runners, scheduled tasks, notifications and CI/CD configuration. CLI output is structured JSON and often includes agentHints.nextSteps; all commands support --help.
Use this skill to verify Compose layout measurements from a UI test using `assertWidthIsEqualTo`, `assertHeightIsEqualTo`, `assertWidthIsAtLeast`, `assertHeightIsAtLeast`, `assertTouchWidthIsEqualTo`, `assertTouchHeightIsEqualTo`, `assertPositionInRootIsEqualTo`, `assertTopPositionInRootIsEqualTo`, `assertLeftPositionInRootIsEqualTo`, plus read helpers `getUnclippedBoundsInRoot`, `getBoundsInRoot`, `getAlignmentLinePosition`, `getFirstLinkBounds`, and the underlying `Dp.assertIsEqualTo(expected, subject, tolerance = Dp(.5f))`. Covers the half-dp default tolerance, the unclipped vs clipped distinction, the canonical "compute padding from two unclipped rects" pattern, and minimum-touch-target assertions like `assertHeightIsAtLeast(MinHeight + 1.dp)`. Use when the developer wants to assert sizes, padding, alignment, position in dp, or asks about `getUnclippedBoundsInRoot`, `DpRect`, touch-target size, or compares widths in pixels. If the developer is comparing layout dimensions from a test, use this skill.
Run technical quality checks across accessibility, performance, theming, responsive design, and anti-patterns. Generates a scored report with P0-P3 severity ratings and actionable plan. Use when the user wants measurable accessibility, performance, responsive, theming, or anti-pattern findings—not when they mainly want an overall UX critique or final visual/detail polish.
Use when the work spans several files or replaces a framework, language, auth, storage, or API layer (rewrites and migrations included), and the user wants it to end up working and tested. Asks only what is missing, writes a spec with acceptance criteria, then runs and verifies it in the background. Not for questions, small edits, or an existing Seed (use run)
Use this skill to pick the correct Compose UI test entry point. Compares `createComposeRule()`, `createAndroidComposeRule<A>()`, `createEmptyComposeRule()`, `runComposeUiTest { }`, and `runAndroidComposeUiTest<A> { }`, plus the v1 vs v2 split (`UnconfinedTestDispatcher` vs `StandardTestDispatcher`). Encodes the rule that mixing `runComposeUiTest { }` and a `ComposeTestRule` in the same test is forbidden, that `setContent` may only run once, and that `createEmptyComposeRule()` returns `ComposeTestRule` (no `setContent`). Use when the user asks "ComposeTestRule vs ComposeUiTest", reports `IllegalStateException: setContent can only be called once`, mentions `runComposeUiTest`, the v2 deprecation warning, `effectContext`, custom `ComponentActivity`, or "host tests need a coroutine scope".
Browser automation via Puppeteer CLI scripts (JSON output). Capabilities: screenshots, PDF generation, web scraping, form automation, network monitoring, performance profiling, JavaScript debugging, headless browsing. Actions: screenshot, scrape, automate, test, profile, monitor, debug browser. Keywords: Puppeteer, headless Chrome, screenshot, PDF, web scraping, form fill, click, navigate, network traffic, performance audit, Lighthouse, console logs, DOM manipulation, element selector, wait, scroll, automation script. Use when: taking screenshots, generating PDFs from web, scraping websites, automating form submissions, monitoring network requests, profiling page performance, debugging JavaScript, testing web UIs.
Evaluate an interface from a UX perspective, assessing hierarchy, information architecture, emotional resonance, cognitive load, and overall quality with evidence-linked judgments and actionable feedback. Use when the user wants an overall design or UX review—not when the main need is measurable accessibility/performance diagnosis, or final micro-detail polish.
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy
Use when Codex is building or iterating on a web game (HTML/JS) and needs a reliable development + testing loop: implement small changes, run a Playwright-based test script with short input bursts and intentional pauses, inspect screenshots/text, and review console errors with render_game_to_text.
Improve slow Django test suites through measured profiling, safe test settings, database and migration tuning, parallel execution, test-data cleanup, and CI workflow changes. Use when Django or pytest-django tests are slow, CI test jobs take too long, database setup dominates runtime, or a user asks for a prioritized speedup plan rather than isolated test fixes.
Review Effect v4 code, including beta, RC, and release versions, for correct error propagation, resource ownership, cancellation, concurrency, primitive selection, services/layers, schemas, atoms, observability, and tests. Use for requests such as "effect v4 review", "review my Effect 4 changes", "check Effect patterns", or "audit the Effect codebase" when the project uses v4. Supports files, working changes, PR/branch diffs, and repository audits. Reviews with one agent by default; delegation has an explicit total limit. Verify APIs against the installed version. For Effect v3 use effect-review.
Use this skill to drive Jetpack Compose text fields from tests with the text-specific actions — performTextInput (insert at cursor via the InsertTextAtCursor semantics action), performTextReplacement (clear + replace via SetText), performTextClearance (replace with empty string), performTextInputSelection (set TextRange selection), and performImeAction (fire the configured IME action such as Next, Done, Search). Covers the auto-focus path, the enabled / editable / focusable preconditions, the append vs replace gotcha that produces "helloworld" instead of "world", and the IME-action focus chain (Next advances focus to the next focusable). Use when the developer asks "how do I type into a TextField in a test", "test an IME Next button", "test a clear-text action", "BasicTextField won't accept input", "text input duplicated in test", or reports "Failed to perform text input" / "Failed to perform IME action" / "Default ImeAction" errors.
Guides developers and testers in writing, fixing, and structuring Gatling load test scenarios. Use this skill whenever the user mentions load testing, performance testing, stress testing, Gatling, virtual users, VUs, ramp-up, injection profiles, simulations, JMeter migration, k6 migration, throughput, response time SLAs, or wants to benchmark an API or service — even if they don't explicitly say 'Gatling' or 'performance test'.
Convert raw notes, error logs, voice dictation, or screenshots into crisp GitHub-flavored markdown issue reports. Use when the user pastes bug info, error messages, or informal descriptions and wants a structured GitHub issue. Supports images/GIFs for visual evidence.
Write property-based tests using Hegel across Rust, Go, C++, TypeScript, Java, and OCaml projects. Use this skill whenever the user asks to write tests, add test coverage, or improve testing for functions, modules, or libraries — especially when the code has properties like round-trips, invariants, or contracts that hold across many inputs. Also triggers on: "property-based tests", "PBT", "hegel", "fuzz", "generative tests", "randomized testing", "test with random inputs", "shrinking", or when existing tests use proptest, quickcheck, rapid, gopter, rapidcheck, fast-check, jqwik, junit-quickcheck, qcheck, or crowbar.
Design ablations that isolate causal mechanisms instead of producing decorative tables. Use during paper preparation, benchmark analysis, reviewer response, or whenever multiple mechanisms could explain an observed gain and the user needs to determine which component matters, why it matters, and whether alternative explanations remain plausible.
Convert operational incident and near-miss evidence into durable product, engineering, test, evaluation, and governance improvements with verified closure. Separate observed facts from causal hypotheses and unresolved uncertainty; map follow-up work across code, tests, skills, operations, product, and governance; track ownership, verification, and closure for every finding. Do not use to assign blame or produce a generic postmortem template; do not close learning because tickets were created — require evidence the intended change occurred.
Use this skill to drive Jetpack Compose UI from tests with non-touch input — performMouseInput (click, rightClick, doubleClick, tripleClick, longClick, animateMoveTo, dragAndDrop, smoothScroll, enter / exit, press / release / scroll), performKeyInput (keyDown, keyUp, isKeyDown, modifier-state vals isCtrlDown / isShiftDown / isAltDown / isMetaDown / isFnDown / isCapsLockOn / isNumLockOn / isScrollLockOn, helpers pressKey, withKeyDown, withKeysDown, withKeyToggled, withKeysToggled), performKeyPress for direct KeyEvent injection, and performMultiModalInput for hover-then-click flows. Covers MouseButton (Primary, Secondary, Tertiary), ScrollWheel (Horizontal, Vertical), and the repeat-key behaviour driven by advanceEventTime. Use when the developer asks "how do I right-click in a Compose test", "test a Ctrl+S shortcut", "send a keyboard shortcut", "simulate hover", "test a tooltip", "scroll the mouse wheel", or "drag and drop with a mouse" on desktop or Compose Multiplatform.
Real Chrome/Brave + macOS Computer Use from inside the browser — zero CDP fingerprint, real sessions; mandatory for visual deploy verification. Drive clickwork yourself. USE WHEN verify deploy, confirm UI, screenshot verification, computer use, macos automation, debug web, troubleshoot, visual check, motion/animation bug, jank, transition stutter, scrub a flow, console logs/errors, runtime/JS/react errors, mismatch warning, network traffic/log, HAR/pcapng export, hydration/blank-page debug, flash then blank/page broken, why is this not working/what's happening on the page, authenticated page, bot detection bypass, reproduce bug, drive native app, about to ask the operator to click/navigate/fill a form/log in/approve OAuth, OAuth consent flow, complete a web login, do this in your browser. NOT FOR residential-proxy crawling (BrightData) or social actor scraping (Apify).
Review Kafka producer and consumer performance configurations in both the live cluster (via Lenses MCP) and the codebase. Flags un-tuned defaults, anti-patterns and missing best practices. Use when user says "review Kafka performance", "check producer configs", "tune Kafka settings" or asks about throughput, batching or compression. Do NOT use for cluster sizing or capacity planning.
Review the current diff for regressions, correctness bugs, tests, simplifications, and docs issues, scaling depth to a low/medium/high/xhigh/max effort level. Use ONLY when explicitly requested by name: the user runs /material-ui-review, writes $material-ui-review, or asks for "the Material UI review skill". Do NOT use for general review requests such as "review my changes", "review this diff/branch/PR", or after finishing an implementation — handle those without this skill unless the user names it. Pass --comment to post a top-level PR comment, --comment inline for inline PR comments, or --fix to apply findings.
Evaluate a model rigorously and inspect where it fails. Use after training, during comparison to baselines, for LLM/system evaluations, noisy or suspiciously clean results, or whenever benchmark numbers alone are insufficient and the user needs slices, failure categories, and explanations.
Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.
Multi-agent review-and-improve loop for a GitHub PR you have checked out: applies fix commits to the local branch, pushes them to the PR, and reports back in a PR comment. Use when the user wants to "improve a PR", "review and commit fixes", "iterate on my PR", or "review and push back" against a checked-out PR branch; the PR is auto-detected when no locator is supplied. Activates the `review-anvil` engine in per_fix mode.
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.