$ skills --search --map --checkout
Every card says what the skill does and what it looks for. Add the ones you want, and the cart pulls in what they depend on. Check out as a zip for Claude Code, Claude.ai, Codex or ChatGPT, with the source, license and changes for each item.
Checks and fixes semantics, keyboard, focus, contrast, labels and motion against WCAG 2.2 AA.
Captures one decision with its context, rejected options and consequences in a numbered, immutable record.
Gives every agent the same baseline rules on scope, reuse, reading order and verification.
Routes each ask to one owning role, then enforces work order → specialist → verify → handoff.
Locks request/response shapes and change rules so clients and servers can ship independently without breaking.
Specifies an append-only audit events model and the rules for writing, reading, retaining and exporting it.
Runs a reproduce → isolate → hypothesize → fix → prove loop so fixes address causes, not symptoms.
Runs five opposing advisor views and a cross-examination, then a chair commits to one call and writes it up.
Gives implementers a working checklist so every screen uses primitives and tokens instead of one-off styles.
Cross-checks docs against code registries, copy and git history and lists each drift with its file.
Updates the docs that a change made stale, documenting only what was actually built.
Turns asks into work orders, sizes them, assigns one owner, and closes them only after verify passes.
Applies a clarify → smallest change → verify → explain loop so changes are small, correct and honest.
Maps stable codes (ERR_/WARN_/INFO_/SUCC_) to titles, messages and severity so UI and API stay consistent.
Converts prose or code into a readable Mermaid diagram with the right diagram type and conventions.
Records done / not done / decisions / files / risks / next 3 actions so work survives a role switch or a pause.
Sets the rules for presets, wrappers, reduced motion and performance so animation is consistent and 60 fps.
Runs a dependency decision tree and recommends stdlib, an existing dep, a default pick or a small local function.
Blocks or clears content for publishing by scanning for secrets, real identities, internal hosts and raw data.
Runs the real checks, records an honest PASS/FAIL/PENDING report in the work order, and never fakes green.
Maps which systems a change touches, what must move together, and records each shipped version in an immutable release table.
Answers "where is X / what does this touch / who calls this" with path:line evidence and stops.
Locks contracts and architecture decisions before specialists build, and unblocks them when the design is unclear.
Checks a screen against platform guidelines and the Laws of UX and names the specific rule each problem breaks.
Reviews only the changed lines of a UI diff and reports blocking vs. should-fix findings with quoted code and fixes.
Reviews layout, colour, density and component boundaries and outputs a precise design brief; does not write production code.
Runs all checks to completion and reports pass/fail per suite with quoted errors, separating code failures from environment problems.
Turns an ask into a scoped work order file with acceptance criteria, story points and verify commands.
Guides distinctive, intentional visual design (aesthetic direction, typography) so UI does not read as a template.
Guides building high-quality MCP servers in Python (FastMCP) or TypeScript, with evaluation.
Creates, edits and benchmarks agent skills, including evals and description tuning for better triggering.
Tests local web apps with Playwright: drives the UI, captures screenshots and reads browser logs.
Builds, debugs and reviews Durable Objects code for persistent state and coordination.
Builds, migrates and deploys Next.js apps on Cloudflare Workers and compares vinext with OpenNext.
Designs, configures and reviews Cloudflare One Zero Trust deployments, including Access-protected apps.
Helps choose Cloudflare products for an app and diagnoses Cloudflare error pages (52x, 1xxx).
Sets up, repairs or migrates to Cloudflare Turnstile bot verification with server-side Siteverify.
Audits and improves page load and interaction performance, Core Web Vitals and Lighthouse scores.
Reviews and writes Cloudflare Workers code against current platform best practices, preferring retrieved docs over memory.
Runs and troubleshoots Wrangler CLI commands for local dev, previews, deploys and resource management.
Automates browser interactions, tests web pages and works with Playwright tests through the Playwright CLI.
General browser automation with Playwright: detects dev servers, writes test scripts, screenshots, responsive and link checks.
Postgres rules for schema, indexes, RLS, migrations and query performance, prioritized by impact.
Covers Supabase Database, Auth, Edge Functions, Realtime, Storage and CLI, checking the changelog before implementing.
Turns a rough idea into a design through structured questions before any code is written.
Executes a written plan in batches with checkpoints for review.
Dispatches a reviewer subagent to check work against the plan before merging.
Runs a plan by dispatching a fresh subagent per task with review between tasks.
A four-phase root-cause debugging process before proposing any fix.
Enforces red-green-refactor: a failing test before any production code.
Creates isolated git worktrees for feature work with safety checks.
Requires running verification commands and showing output before claiming work is done.
Writes detailed, step-by-step implementation plans that another agent can execute.
Audits GitHub Actions workflows where AI agents run in CI for prompt and expression injection paths.
Builds CodeQL databases and runs security queries over a codebase.
Security-focused review of a PR or diff with blast-radius, git-blame context and test coverage, written up as a report.
Plans and runs Semgrep security scans with approved rulesets and merges results to SARIF.
Finds error-prone APIs, dangerous configurations and footgun designs that invite security mistakes.
Audits dependencies for advisories, abandoned upstreams, publisher concentration and install-time scripts.
React composition patterns: compound components, context providers and avoiding boolean-prop sprawl.
React and Next.js performance rules from Vercel: data fetching, bundle size, rendering and re-renders.
Reviews UI code against Vercel's Web Interface Guidelines for UX and accessibility.