Claude Skills by its-magdy

github.com/its-magdy

37 skillsA× 33B× 3F× 10 installs2 views

ApproveARecord human approval of an ideated case checklist — mints stable row ids and writes the > HUMAN APPROVAL banner reviewer Check 14 (and the planner) read. PRECONDITION — a `.cases.md` must exist (run /qa-warden:ideate first).testingexpressgitVotes: 0GitHub stars: 2
Axe A11yACompile an `a11y_violations_below` oracle into a settled, full-page @axe-core/playwright WCAG scan, or run the a11y half of a `/qa-warden:review url=` live-page audit. Use when a spec's oracle declares that key or when auditing a live page. Not for specs whose oracle does not declare it — the scan would be an orphan assert.testinggotestingVotes: 0GitHub stars: 2
Batch FixAApply a single healer pattern fix across every failing test matching a path-or-substring filter.testingrustshellVotes: 0GitHub stars: 2
CoverageAHonest test-coverage report across 9 static dimensions (route footprint, requirement, spec-without-test, assertion, lens, flow, case, waiver destination, a11y need→oracle), from basis, specs, cases and route manifests. No test run, no line coverage, WARN-only. Use when the user asks what is or isn't covered, where the test gaps are, or how complete the suite is.testingshellbashVotes: 0GitHub stars: 2
DoctorARead-only scripted health check of a scaffolded QA project — last-run freshness and zero-test guards, substrate drift, vocab mirrors, stale context and manifests, orphan sentinels, unreviewed or uncompiled specs. Use when something looks off, before trusting a green run, after a plugin upgrade, or when another command points here. Writes nothing; `--verify-invariants <spec>` is the only mode that runs tests.testingrustgoVotes: 0GitHub stars: 2
ExploreAInvoke the exploration subagent — hot mode (refresh app.context.md) or area mode (write specs/_context/<site>/<area>.md). Default = hot ONLY on a bare invocation; an area invocation missing mode=/site=/area= STOPs rather than defaulting. Use when a site's auth/env has changed, an area's context file is missing or past its staleness tier, or planner/healer reports stale/missing context and hands off here.testingbashapiVotes: 0GitHub stars: 2
Skills DirectorySkills Directory Look up any of 103,619 scanned skills — grade, score, and every finding. Free.Check a skill
GenACompile a Markdown+YAML spec into a Playwright .spec.ts via the generator subagent.testingrustbashVotes: 0GitHub stars: 2
HealATriage a failing Playwright test via the healer subagent; patch selectors/waits or file a bug.testinggoshellVotes: 0GitHub stars: 2
HelpAAnswers any question about the QA Warden plugin — how it works, what a command, agent or skill does, what to do next, or how to reach a goal — using the project's current state to give a situated answer.testingrustgoVotes: 0GitHub stars: 2
IdeateAEnumerate a CHECKLIST of candidate test cases for a feature/bug/enhancement via the ideation subagent (SFDIPOT lens fan-out + de-dup + risk-rank + completeness critic). Reads the .basis.md; produces .cases.md for you to approve/prune. Does NOT write specs or tests. PRECONDITION — a `.basis.md` must exist (run /qa-warden:intake first).testinggoVotes: 0GitHub stars: 2
ImpactATest-impact analysis — lists every spec affected by a change to a route, field, factory, area, GraphQL operation or business source, from artifacts/route-manifests/ (and specs/_context for source=). Use when the user asks which tests a change could break, what to re-run after a change, or what covers a route or field.testingbashVotes: 0GitHub stars: 2
Import CasesAImport existing manual test cases (TestRail/Zephyr/Xray CSV, markdown, or pasted table) into the plugin's cases.md + a minimal basis — provenance-stamped, dedupedtestinggoc#Votes: 0GitHub stars: 2
InitFBootstrap the QA runtime substrate into the current project (configs, scripts, fixtures, CLAUDE.md, permissions) and install dependencies. Run once per project after installing the QA Warden plugin.testingrustbashVotes: 0GitHub stars: 2
IntakeAInteractively capture the "test basis" for a feature/enhancement/bug — auto-gather observable context, then interview only for non-observable intent (the oracle) — and write specs/_context/<site>/<area>/<feature>.basis.md. First step of the authoring chain (intake → ideate → approve → new-spec); PRECONDITION — the area's context (app.context.md + <site>/<area>.md) MUST exist; intake hard-stops and tells you to run /qa-warden:explore first if it doesn't.testingrustgoVotes: 0GitHub stars: 2
Metamorphic RelationsAGenerate 2-3 metamorphic-relation "twin" specs of a passing `.spec.ts` — invariant-preserving transforms whose outcome must match the parent, run nightly to catch specification drift. The verifier authors them (Process step 8a) right after a new spec's first green run; the read-only reviewer only checks that they exist and agree.testingrustgoVotes: 0GitHub stars: 2
New SpecADraft a new Markdown+YAML spec for one feature via the planner subagent. Runs LAST in the authoring chain (intake → ideate → approve → new-spec) — PRECONDITIONS — the area's context (specs/_context/app.context.md + <site>/<area>.md) must exist and be fresh, and if a .cases.md exists it must be APPROVED. The planner STOPs (handoff) on missing/stale context or un-approved cases.testinggoshellVotes: 0GitHub stars: 2
Playwright CliADrive a real Chromium/Firefox/WebKit page deterministically from an agent. Use when authoring or healing Playwright tests, running a single navigate-snapshot-assert loop, or any healer/nightly flow that must stay token-predictable. Prefer this CLI over Playwright MCP for healer, generator, and nightly; reach for MCP only for first-contact exploration.testinggoshellVotes: 0GitHub stars: 2
ReportAAggregate artifacts/last-run.json + reports/*.json into a PR/Slack-ready summary — totals, failure-signature grouping, durations, flake/retry surfacing, parked defects, value ledger. Use at the end of a run-of-record suite run (CI, or QA_RUN_OF_RECORD=1), or ad-hoc when asked "what did the suite just do?"testingrustgoVotes: 0GitHub stars: 2
RetireACoherently retire a feature — dry-run inventory of all linked artifacts, consumer-checked deletion, tsc + list verificationtestingrustshellVotes: 0GitHub stars: 2
ReviewARuns the load-bearing reviewer subagent over the current diff or a named spec/path — the full assertion-contract check suite. With url=<x>, audits a live URL instead (a11y + visual + closed vocab). Use before committing or merging a test, or when the user asks whether a test can actually fail.testingrustgoVotes: 0GitHub stars: 2
RunAExecute Playwright tests in one of four ways — run a single spec headless/CI-shaped (mode=single), run the @smoke-tagged suite (mode=smoke), repeat-probe a spec N times to check flakiness/repeatability/stability (mode=repeat), or run only the tests a git change touches (mode=changed). Deterministic, JSON-only, no LLM in the loop. Do not hand-roll a bare `npx playwright test` for any of these — this skill owns the resolver, freshness/zero-test guards, and report paths every downstream consumer...testinggoshellVotes: 0GitHub stars: 2
Test Case ContractASpec format contract for specs/**/*.md: the YAML schema, the closed 16-key oracle vocabulary and argument shapes, fault:/clock: situation steps. Load before writing, compiling or reviewing a spec.testingrustgoVotes: 0GitHub stars: 2
Test Data SeedABuild worker-scoped, API-seeded, teardown-by-tag test fixtures that are parallel-safe at 10+ workers. Use whenever a spec needs a user, order, org, or any writeable resource — UI-seeded and hardcoded-email patterns collide immediately at parallel scale. Invoked during generator authoring or when a flake post-mortem traces to data collisions across workers.testingtypescriptgoVotes: 0GitHub stars: 2
Visual RegressionAAdd Playwright `toHaveScreenshot()` assertions for layout/CSS/typography regressions that text oracles miss. Use in `/qa-warden:review url=` live-page audits and when a human hand-hardens a visually meaningful flow outside the managed spec contract. Not for generated specs — a screenshot expect has no oracle key, so reviewer Check 2 flags it as an orphan.testinggobashVotes: 0GitHub stars: 2
AnalyzeAUnderstand failures and improve the eval assets: cluster failures, label judge-calibration cases (15-minute flow), edit rubrics while grading, mine traces for new cases, and manage the sealed holdout. Use with --transcripts when the user has exported chats, support logs or saved traces and wants to know what is going wrong — that branch needs NO eval setup: no profile, no cases, no run. Otherwise use after runs to triage results, to calibrate the judge, or to promote real conversations into t...ai-agentspythongoVotes: 0GitHub stars: 2
DiscoverAProfile a target LLM app: read its code and/or traces, determine architecture (agents, route targets, tools, flow), write .evalup/profile.yaml and adapter.yaml, patch missing instrumentation, and report design findings. Use when pointing evalup at an app for the first time, or with --diff after the app changed to find stale eval cases.ai-agentspythongoVotes: 0GitHub stars: 2
GenerateAGenerate an eval dataset for the profiled app: coverage-gridded cases with expected routes/trajectories/rubrics, splits (full plus its smoke and canary subsets and a sealed holdout), and a targeted human-review pass. Use when the app has a verified profile and needs test cases, or to expand coverage for a specific layer or domain.ai-agentspythongoVotes: 0GitHub stars: 2
HelpAExplain the evalup plugin and point the user at the one command or document that answers them: what evalup is and what it measures, how the workflow runs from first contact to steady state, where to start or resume, which of the commands fits the situation they describe, what the modes and flags mean, who has to be involved (developer, reviewer/QA, domain arbiter) and whether they need Claude Code, where eval state lives, what the scores can and cannot claim, and why something is locked, gate...ai-agentspythongoVotes: 0GitHub stars: 2
OptimizeAImprove the app's prompts and tool descriptions through a failure-driven loop: reflect on failure clusters, propose one targeted edit, measure on the training split, confirm on the sealed holdout with a statistical gate, keep or revert. Use when eval scores have plateaued or a failure cluster implicates a specific prompt or tool description.ai-agentspythonshellVotes: 0GitHub stars: 2
RunAResolve a run mode into a plan, execute it with run_cases.py, and turn the result into a report. The runner owns pre-flight, every app call, every scorer, the artifacts and the baseline diff; this skill decides what to run and what the numbers mean. Five modes — smoke, regression, targeted, holdout, full — cover everyday edits through release validation, plus a CI gate.ai-agentspythonrustVotes: 0GitHub stars: 2
StartAGuided entry point for evalup: detects the current state of a target app's eval setup (profile.yaml, datasets, pinned baseline) and routes the user to exactly the next step — discover, generate, run, analyze, or optimize (which it names; the user invokes that one) — telling them where they are ("step 2 of 5"). Use this whenever the user wants eval work DONE on an app: to begin or resume evaluating it, to set up evalup, to get test cases or a baseline for an agent or chatbot, to find out what ...ai-agentspythongoVotes: 0GitHub stars: 2
Audit DesignAReviews a Figma export the way a senior frontend / iOS / Android engineer does before any code is written — token and typography binding, spacing grid, sizing and responsiveness, component variants and interaction states (hover/pressed/focus/disabled/error), loading/empty/error screens, content edge cases, touch targets, contrast, font scaling, dark mode, RTL, platform chrome and safe areas, assets and effects that won't translate — and turns every gap into a concrete question for the designe...designgoswiftVotes: 0GitHub stars: 2
Build ScreenABuilds a front-end screen as real, production-quality code from a Figma export, for any stack (React/Tailwind, CSS Modules, React Native, SwiftUI, Jetpack Compose, Flutter, or a custom target). Use this whenever the user wants a design turned into code — "implement the login screen", "build this frame", "code up the settings page", "make this design real", "match the Figma" — or points at a design/ export or a Figma frame, even if they don't say which stack. Resolves the stack, audits the des...designgoswiftVotes: 0GitHub stars: 2
ExtractBGets a design out of Figma and onto disk as JSON + assets, using the local "Design Twin" plugin (free Figma plan, no Figma API, no network egress). Use this whenever the user wants design data pulled, exported, refreshed or re-pulled from Figma — "pull the login screen", "export this frame", "get the design system", "get me the colors", "set up my theme from Figma", "grab the components from Figma" — and also whenever a build/codegen task needs files that aren't in design/export/ yet, or the ...designgoswiftVotes: 0GitHub stars: 2
HelpBOrientation, connection, one-time setup and troubleshooting for Design Twin, the free-plan Figma→code workflow. Use this whenever the user asks a discovery-shaped question — "how does this work", "what can this plugin do", "how do I connect Claude to Figma", "is there a CLI", "what's the MCP server for", "how do I set this up", which export path to use — even if they don't name a command; when they are about to try a Figma task but have no connection set up yet; and whenever something is brok...designgoswiftVotes: 0GitHub stars: 2
Sync DesignAUpdates already-built code after the Figma design changed, by patching only what moved instead of rebuilding the screen. Use this whenever a screen that exists in code needs to catch up with its design — "the designer changed the header", "the design was updated, update my code", "re-sync this screen with Figma", "what changed in the design since I built it?", "the primary color changed", "apply the new spacing" — or when the user re-pulled an export and asks what to do next. It snapshots the...designrustbashVotes: 0GitHub stars: 2
VerifyBChecks an already-built screen against its Figma design, without changing any code. Use this whenever the user asks whether code matches the design — "does this match the design?", "check the login screen against Figma", "compare my build to the mockup", "is this pixel-accurate?", "what's off in this screen?", "verify the settings page" — for a screen that exists in code and has an export under design/export/. An agent renders the screen, measures every visible node and drives every designed ...designrustgoVotes: 0GitHub stars: 2