yonatangross/orchestkit

AI UI Generation covers prompt patterns, tool selection, review checklists, token injection, and CI gates for AI-produced UI. Use it when generating components or full-stack UI with json-render, v0.app, Stitch, Bolt Cloud, or Cursor.

API Design covers REST and GraphQL contract design, versioning schemes, RFC 9457 Problem Details, and OpenAPI specs. Use it when defining endpoint shapes, deprecation windows, or standardizing error responses across services.

Accessibility provides patterns for WCAG 2.2 AA compliance, keyboard focus management, React Aria components, and honoring user preferences. Use it when building screen-reader support, keyboard navigation, dialogs, forms, or reduced-motion and contrast handling.

Activation Audit checks real spawn telemetry to see which OrchestKit sub-agents are being used, which never fire, and how often generic agents replace specialists. Use it before pruning the catalog or after changing agent spawn paths.

Agent Orchestration provides patterns for agent loops, multi-agent coordination, framework selection, and multi-scenario workflows. Use it when building autonomous loops, routing work across agents, or comparing orchestration frameworks.

Analytics queries local OrchestKit data to report agent usage, skill frequency, hook timing, team activity, session replays, cost estimates, and trends. Use it when reviewing performance or understanding usage patterns from privacy-safe local files.

Animation And Motion Design covers Motion and the View Transitions API for component animations, page transitions, gestures, and exit effects. Use it when building interactive UIs or adding reduced-motion support.

Architecture Decision Records provides a Nygard-format template for documenting architectural decisions with context, decision, consequences, and alternatives. Use it when recording important technology choices or evaluating trade-offs.

Architecture Patterns validates clean architecture, backend layer separation, project structure, test standards, and architecture sizing. Use it when reviewing boundaries, folder conventions, dependency direction, or whether a project is over-engineered.

Assess scores code, designs, or architectures across correctness, maintainability, security, performance, testability, and simplicity. It produces pros/cons analysis, compares against project conventions, and suggests actionable improvements.

Async Jobs covers background task processing patterns for Celery, ARQ, Redis, and Temporal. Use it when implementing queues, retries, scheduling, routing, or distributed task execution.

Auto classifies a plain-English goal and routes it to the right OrchestKit specialist skill. Use it as the default entry point for goal-shaped requests unless you are already inside another skill.

Bare Eval runs `claude -p --bare` for isolated evaluation, trigger testing, and LLM grading without plugin or hook interference. Use it when benchmarking prompts, scoring outputs, or checking skill matches in a clean non-interactive call.

Brainstorm designs ideas into structured alternatives using parallel agents, feasibility checks, and scoring. Use it when exploring solutions, comparing approaches, or iterating on a design problem.

Browser Tools adds safety guardrails to the upstream agent-browser workflow, including URL blocklisting, rate limiting, robots.txt enforcement, and scraping limits. Use it when browser automation needs controlled access and safer crawling.

Business Case helps calculate ROI, NPV, IRR, payback period, and TCO for investment decisions. Use it for build-vs-buy comparisons, cost-benefit analysis, and sensitivity testing.

CI Debug classifies a failing CI run using an 11-pattern playbook and identifies the most likely fix. Use it for a specific PR check or GitHub Actions job that failed, not for broader CI sweeps or app test failures.

Chain Patterns provides patterns for multi-phase Claude Code pipelines, including MCP capability detection, handoff files, checkpoint-resume, worktree-isolated agents, and cron monitoring. Use it when building or debugging a pipeline skill.

Ci Sentinel runs headless CI classification on open PRs with failing required checks, posts a collapsed verdict comment, and logs results per repo. It is used to triage repeated CI failures without auto-applying fixes.

Code Review Playbook provides structured checklists, conventional comments, and feedback templates for reviewing pull requests and standardizing review practice. Use it when doing PR reviews, self-reviews, or team review training.

Competitive Analysis uses Porter’s Five Forces, SWOT, and competitor profiles to assess industry attractiveness and compare rivals. Use it when evaluating market position, threats, feature gaps, or building sales battlecards.

Component Search finds production-ready React components in the 21st.dev registry from natural-language queries. Use it to look up UI building blocks, compare alternatives, and get install instructions.

Full Codebase Audit performs single-pass whole-project analysis to find security, architecture, and dependency issues across files. Use it before releases or security reviews when you need severity-ranked findings from the full codebase.

OrchestKit Configuration opens an interactive wizard for setting MCP servers, hook permissions, keybindings, and install presets. Use it to customize OrchestKit behavior or manage settings for a project or user profile.

Smart Commit creates validated Conventional Commits from staged changes, with scope detection, agent attribution, and hook-compliant messaging. Use it when you want to commit work or generate a commit message safely.