Reviews REST APIs by linting the OpenAPI spec, detecting breaking changes, and grading design quality.
An installation skill that wires a local Ollama MCP server into NanoClaw so the container agent can run local models and optionally manage the model library.
Copies, registers and validates the Cisco Webex adapter for a NanoClaw project via the Chat SDK bridge.
Copies, registers and verifies the Resend email adapter for NanoClaw, then wires your own address as owner and sends a hello email.
A measurement-first Go skill covering benchmark authoring, pprof profiling, benchstat statistical comparison, and CI regression detection.
Designs startup-grade AWS architectures and produces CloudFormation/CDK/Terraform code plus cost-optimization plans.
Re-runs each installed NanoClaw skill's own idempotent apply step so their code is refreshed from the upstream branches.
Generates Angular code and architectural guidance following official style guides for signals, forms, routing, DI, and testing.
A rulebook that makes Claude write Next.js 15 + React 19 + Shadcn/ui code following consistent Server Component and App Router best practices.
Searches and applies up-to-date web platform best-practice guides before any HTML/CSS/client-side JS work.
A four-mode code quality skill covering technical-debt cleanup, Ruff/Biome linting, a Python quality gate, and a multi-language universal gate.
Verify UI, console, network, performance and accessibility in a real browser through the Chrome DevTools MCP server.
Trains Claude to ship features as thin, working vertical slices — implement, test, verify, commit — instead of dumping an entire feature at once.
Debug multi-cause bugs by splitting them into competing hypotheses, investigating them in parallel, and arbitrating the true root cause from evidence.
A skill that fetches live Vercel AI SDK docs and AI Gateway model data so Claude codes against the current API instead of stale memory.
Reviews GitHub pull requests with six parallel specialist subagents, adversarial verification, and confidence/severity gates so only high-signal findings get posted.
Delegates coding, code review, and plan review to the OpenAI Codex CLI with strict sandbox and review-safety rules.
An end-to-end workflow that turns a GitHub issue into a tested fix and a submitted pull request.
Generates Conventional Commits messages with ≤50-character subjects that explain the "why" instead of restating the diff.
A fan-out audit pipeline that runs every analysis skill in parallel and produces a unified health dashboard in Markdown and HTML.