yonatangross/orchestkit-plugin

analytics

Local OrchestKit usage analytics: agent and skill frequency, hook timing, session replay, cost estimates, model delegation trends. Use when reviewing usage, performance or costs.

api-design

API wire contracts for REST and GraphQL: resource shape, versioning and deprecation, RFC 9457 errors, OpenAPI. Use when designing an endpoint contract. Not for runtime code (use python-backend).

architecture-decision-record

Nygard-format ADRs: context, decision, consequences, alternatives. Use when writing an ADR or recording an architectural decision.

architecture-patterns

Clean and layered architecture patterns, project structure and test-standard validation. Use when designing system boundaries or enforcing structure. For tier sizing, use scope-appropriate-architecture.

assess

Rates code, designs or approaches 0-10 across quality dimensions with pros, cons and fixes. Use when evaluating or comparing options. Not for fixing or building (use implement or fix-issue), PR review (use review-pr) or running tests (use verify).

audit-activation

Audits sub-agent activation from telemetry: attempted, started, completed, unattributed spawns, and zero-start agents. Use when agents feel under-used or before pruning the catalog.

audit-full

Whole-codebase audit in one 1M-token pass: security, architecture and dependencies with severity-ranked findings. Use before a release or security review. Too big for one pass: use audit-full-mapreduce.

auto

Default front door for any goal-shaped request: classifies a plain-English goal and routes it to the right skill. Use for do this, figure out, I want, help me, fix, build, improve. Skip only inside another skill (no recursion).

bare-eval

Isolated eval and grading calls with claude -p in bare mode, free of plugin and hook interference. Use when grading skill outputs, benchmarking prompts or testing trigger accuracy.

brainstorm

Design ideation with parallel agents: 10+ ideas, feasibility filter, devil's-advocate scoring, trade-offs. Use when brainstorming or comparing alternatives. Not for building or fixing code (use implement or fix-issue) or scoring a design (use assess).

browser-tools

Safety rules over the upstream agent-browser skill: scraping ethics, auth, request pacing, local URLs. Use for browser automation.

business-case

ROI, NPV, IRR, payback and TCO for investment decisions. Use when building a financial justification, cost-benefit or build-vs-buy analysis.

chain-patterns

Multi-phase pipeline patterns: MCP detection, handoff files, checkpoint-resume, worktree agents, cron monitoring. Use when building or debugging a pipeline skill.

ci-debug

Diagnoses one failing CI run or PR check against an 11-pattern playbook and proposes the exact fix, never applying it unasked. Use when a GitHub Actions job failed. Not for org-wide sweeps (use status) or app test failures.

ci-sentinel

Daily headless classifier for every open PR with red required checks: runs ci-debug, posts a collapsed verdict comment, logs a ledger. Proposes only, never pushes. Use to automate recurring CI triage.

code-review-playbook

Review process, conventional comments, language checklists and feedback templates. Use when reviewing code or standardizing review practice. To review a PR end to end, use review-pr.

commit

Creates Conventional Commits with scope detection, co-author lines and hook compliance; blocks secrets and generated-only files. Use when asked to commit, save progress or write a commit message, not for incidental commits.

competitive-analysis

Porter's Five Forces, SWOT, competitor profiles, feature gaps and battlecards. Use when analyzing market position or building sales battlecards.

component-search

Searches the 21st.dev registry for production React components, ranked with install steps. Use when looking for a UI component or an alternative to one.

configure

Interactive wizard for OrchestKit settings: MCP servers, hook permissions, keybindings, presets. Use when customizing plugin behavior.

cover

Writes tests that do not exist yet across unit, integration and end-to-end tiers, healing failures up to 3 times. Use when code lacks tests. Not for grading existing tests (use verify) or running a suite.

create-plan

Writes an implementation plan file (.ork/tasks/TASK/NN-plan-SLUG.md) with phases, file changes and runnable success criteria. Use when research is done and a reviewable plan is needed before code.

create-pr

Opens a GitHub PR after pre-flight checks (tests, lint, types, security) with a conventional title and summary. Use only when asked for this skill; an everyday gh pr create stays plain tooling.

create-research

Documents how code works today for a task's questions, with file:line citations. Use before planning a change; describes, never proposes. NOT for mapping a repo (use explore).

database-patterns

Schema design, Alembic migrations and database versioning. Use when creating migrations, designing or normalizing schemas, or fixing schema drift.

demo-producer

Demo videos for skills, plugins, CLIs and code walkthroughs: scripts, storyboards, VHS terminal recordings, Remotion. Use when producing a showcase or marketing video.

design-context-extract

Extracts design DNA from screenshots, URLs or recordings: colors, type, spacing, components and motion as tokens or Tailwind config. Use when asked to extract or audit a design or match a visual identity.

design-import

Scaffolds React components from a Claude Design bundle and stops at files on disk. Use when handed a claude.ai/design URL or bundle. For a full PR with tests and browser checks, use design-ship.

design-ship

Turns a claude.ai/design link into a PR: design-import, then cover, then expect, then open the PR. Use when a design should come back as a PR. For files on disk only, use design-import.

design-stylecards

Catalog of named aesthetic recipes: shadow stacks, glass, gradients, type scales as copy-paste values. Use for polished elevation, glassmorphism, mesh backgrounds or any make-it-look-like-X ask.

design-system-tokens

W3C design tokens: global, alias and component tiers, OKLCH color, Style Dictionary, dark mode. Use when creating token files or theme systems.

design-to-code

Mockup to React component via Stitch, 21st.dev and Storybook MCP, reusing existing components first. Use when building UI from a mockup or screenshot. To call Storybook MCP tools alone, use storybook-mcp-integration.

dev

One-command dev loop: portless HTTPS subdomain, emulate API mocks, dev server and an agent-browser session keyed to the git branch. Use when starting a branch, switching worktrees or resuming a project.

devops-deployment

CI/CD, containers and infrastructure as code: GitHub Actions, Docker, Kubernetes, Helm, Terraform. Use when setting up pipelines or deployments.

distributed-systems

Distributed systems patterns for locking, resilience, idempotency, and rate limiting. Use when implementing distributed locks, circuit breakers, retry policies, idempotency keys, token bucket rate limiters, or fault tolerance patterns.

doctor

Diagnoses OrchestKit and project health: manifests, hooks, skills, agents, MCP servers, CC version and permissions, with severity and remedies. Use to run doctor checks, diagnose issues or check project health before a release.

dream

Nightly memory consolidation: prunes stale entries, merges duplicates, resolves contradictions, rebuilds MEMORY.md. Use when memory needs cleanup. To store use remember; to search use memory.

emulate-seed

Seed configs for Vercel emulate: stateful GitHub, Vercel, Google, Slack, Stripe, AWS and more, not mocks. Use when setting up test environments, CI or offline development.

error-analysis

Clusters real LLM traces (Langfuse or JSONL) into a confirmed failure taxonomy with counts, then proposes pass/fail evals. Use to learn what to measure before writing evals. Not for CI failures.

errors

Claude Code session error triage: categorizes network, auth, model, tool and permission errors, finds prior occurrences, suggests recovery. Use when a session hits errors or failures.