Commands Cheat Sheet
Comprehensive list of all available commands
/command and let AI do the heavy lifting! No coding required, just tell ClaudeKit what you want. Example
/ck:cook "add a login button" → AI builds the entire feature for you! 🎉 Quick reference if you're coming from an older version
All Commands → Skills: All legacy commands migrated to skills. You can use either /cook or /ck:cook, but using the /ck: prefix is recommended to avoid conflicts with Claude Code built-in commands.
Migration Reference
| Old Command | New Command |
|---|---|
/debug | /ck:debug |
/plan | /ck:plan |
/code @plan.md | /ck:cook @plan.md |
/code:no-test | /ck:cook add footer --no-test |
/code:parallel | /ck:cook refactor api --parallel |
/code:auto | /ck:cook add pagination --auto |
/plan:fast | /ck:plan --fast add auth |
/plan:hard | /ck:plan --hard migrate to microservices |
/plan:archive | /ck:plan archive |
/plan:ci | /ck:fix CI build failing --auto |
/fix:ci | /ck:fix deploy pipeline error --auto |
/fix:test | /ck:fix auth tests failing --review |
/fix:types | /ck:fix type errors in utils --quick |
/fix:ui | /ck:fix layout broken on mobile --parallel |
/git:cm | /ck:git cm |
/git:cp | /ck:git cp |
/git:pr | /ck:git pr |
/git:merge | /ck:git merge |
/design:video | /ck:remotion [video or component] |
/design:3d | /ck:threejs rotating globe with markers |
/design:screenshot | /ck:frontend-design |
/design:describe | /ck:frontend-design |
/review:codebase | /ck:code-review codebase |
/review:codebase:parallel | /ck:code-review codebase parallel |
/docs:init | /ck:docs init |
/docs:update | /ck:docs update |
/docs:summarize | /ck:docs summarize |
/content:blog | /ck:copywriting blog [context] |
/content:landing | /ck:copywriting landing [context] |
/skill:create foo | /ck:skill-creator foo |
/integrate:stripe | /ck:payment-integration stripe checkout |
/integrate:sepay | /ck:payment-integration sepay webhook |
/bootstrap:auto | /ck:bootstrap --auto |
/bootstrap:auto:fast | /ck:bootstrap --fast |
/bootstrap:auto:parallel | /ck:bootstrap --parallel |
/test:ui | /ck:test ui [url] |
Claude Code v2.1.1+: Skills invocable via /skill-name [args] or /ck:skill-name — both work identically. Prefer /ck: prefix for skills that conflict with Claude Code built-in slash commands (e.g. /ck:plan instead of /plan).
Understanding the Symbols
Each command card breaks syntax into small pieces: the command name, a subcommand, optional switches, and the parts you fill in yourself.
How to read command syntax
/<command> [subcommand] [your request] --flag Start with the full command name, such as /ck:plan or /ckm:seo. If the syntax includes a subcommand, put it after the command name. Replace [] with your real request, then add any flag you need.
/ck:plan archive [task] [] or <> input --fast --... flags when you want to change how it runs Not every card has all 4 parts. If a card only shows a command name and flags, use the command name first, then add the flag. [] and <> are guide markers only. When using the command, replace them with concrete content and do not keep the markers.
/ck:plan The first part of the command. Copy this name when you want to run that workflow.
archive Put this after the command name to choose a specific job, such as archive, validate, or red-team.
--fast Starts with -- . Add a flag when you want a faster, deeper, or more automatic run.
[search-target] Text in [] marks an input slot. When using the command, replace it with the concrete file, topic, question, or target; do not keep the [] .
<template> Text in <> is a required value. When using the command, replace it with a concrete value; do not keep the <> .
/ckm:plan:* * means the card groups several variants. When using the command, pick one concrete variant.
Chip colors on cards
subcommand Purple chips are subcommands or modes under the command. argument Gray chips are user input slots or examples to replace. --flag Teal chips are optional flags you can add after the command. /ck:plan refactor the checkout page to reduce payment errors --deep Use when the task needs a deeper implementation plan before coding.
/ck:fix mobile menu overlaps content --quick Use when there is a concrete broken behavior to diagnose and repair.
/ck:git cp Use after reviewing the diff when the change is ready to commit and push.
Here, /ck:plan is the command, the checkout refactor is the real request, and --deep is the chosen flag.
⚡ = Token Consumption - Estimated context size when skill is activated. For reference only.
Weight-based means the token cost scales with the skill prompt size. A ⚡ skill loads <5K tokens into context, while ⚡⚡⚡⚡ loads 50K-100K and reduces the remaining conversation budget.
All Commands by Category
Beta Preview
New skills and enhancements in Engineer Kit Beta
/ck:coding-agent-orchestration new Orchestrate multiple coding agents in one workflow
New skill: picks the right agent (Claude Code, Codex, Gemini, OpenCode, Cursor, Amp, Droid, Antigravity) for each stage, splits work safely, and keeps one final integrator.--single/--sequential/--parallel/--review-loop shape the collaboration model; evidence gates before completion.
[task or workflow] --single --sequential --parallel --review-loop /ck:advise new Trusted advisor interview before you plan or build
New skill: reframes a raw idea, issue URL, or spec into exact requirements and goals, then returns honest unfiltered advice — never implements.--html/--md/--wiki/--github ship reports; --agent runs autonomous single-shot counsel (kongming-style).
[prompt-or-url] --html --md --wiki --github --agent /ck:fable-thinking new Fable 5 reasoning protocol for any model
New skill: multi-hypothesis diagnosis, concrete simulation, adversarial self-review, and calibrated claims — removes fluent-but-wrong failure modes.
Use for debugging, architecture trade-offs, contested claims, constrained writing, and high-stakes analysis.
[task or question] /ck:issue-to-plan new Issue → scout → gate → plan branch (no implement)
New skill: hard audit/brainstorm gate before any planning; only on pass: validate, red-team, push plan branch, issue handoff.
Planning-only — stops before cook/ship/PR. Flags: --repo, --plan-ready-label, --decision-label.
<github-issue-url | issue-number> --repo --plan-ready-label --decision-label /ck:goal-warmup new Outcome Contract preflight before long /goal runs
New skill: locks outcome, surfaces blockers, reviews scope drift; ends Ready / Blocked / Decision required.
Never auto-starts /goal. --fast for a lighter pass.
"<goal>" --fast /ck:codex-goal new Codex /goal with a verifiable stop condition
New skill: Codex-native durable objectives with a clear stop condition and validation loop.
Not a safety boundary or unbounded backlog runner — pairs with /ck:goal-warmup for preflight.
<objective | goal draft> /ck:handoff new Session handoff for a fresh agent
New skill: compact, redacted conversation state so the next agent continues without rediscovery.
Preserves decisions and blockers — not a command checklist. Distinct from git-derived /ck:watzup.
[next-session focus] /ck:interview-docs new User-interview docs (README, ADR, strategy)
New skill: guided interview turns the user's own vision into durable docs.
Content comes from answers — not AI invention or code scanning.
<vision | document-path | topic> /ck:folder-context new Subfolder CLAUDE.md + AGENTS.md
New skill: durable local conventions for one subfolder.
Never touches root CLAUDE.md / AGENTS.md.
<target-folder> /ck:research-prompt new Write the research brief, not the research
New skill: one self-contained paragraph a human or AI researcher can execute without follow-up.
Decision-ready framing only — does not run the research.
<research topic | decision> /ck:document-skills new Office doc router (docx / pdf / pptx / xlsx)
New router skill: when document type is unclear, dispatches to the right CK document skill.
Routes to /ck:docx, /ck:pdf, /ck:pptx, or /ck:xlsx.
[document task] /ck:review-pr enhanced New --merge: merge a ready PR and watch CI
Beta adds --merge after review: merge a ready GitHub PR via gh, then watch post-merge CI to green.
GitHub-only surface again (no glab). Compose with --fix / --reply. Stable still documents --fix + --reply only.
<PR number or URL> --fix --reply --merge /ck:cook enhanced --advice supervision + --auto means all steps
New composable --advice runs cook under kongming advisory supervision (pass-through on ship handoff).--auto now auto-approves all steps (stable still: low-risk only + high-risk stop). Plus brainstorm-first hard gate before plan/implement.
--advice --auto Start Here (Essentials)
Start with a basic workflow: Learn → Level → Ask → Brainstorm → Plan → Execute
/ck:ck-help deprecated ClaudeKit usage guide
Explore commands, skills, and workflows interactively.
/ck:coding-level Set coding experience level (0-5)
Customize explanations from beginner to expert.
/ck:ask Answer technical and architecture questions
Expert consultation for technology decisions.
[question] /ck:brainstorm Brainstorm with problem-first inversion + scout-first gates
Mandatory codebase scout before any question; inverts a preselected solution to name the underlying problem first. Brutal-honest tradeoff analysis. --html writes an editorial-magazine report next to the markdown; --wiki publishes outputs to AgentWiki.
--html --wiki /ck:plan Create detailed plans with HTML, GitHub, and AgentWiki output
Modes --hard / --deep / --parallel / --two auto-run red-team; --deep also auto-runs validate. --html writes a self-contained interactive plan.html; --github creates or updates a GitHub issue with branch, summary, and acceptance criteria; --wiki publishes the plan to AgentWiki. Subcommands: archive, red-team, validate.
archive red-team validate --auto --fast --hard --parallel --two --no-tasks --deep --tdd --html --github --wiki /ck:cook All-in-one feature implementation with hard gates
Scout-first → exact-requirements → plan → implement → test → no-side-effects verify → review. Gates block unsafe completion. --no-test override available.
--interactive --fast --parallel --auto --no-test --tdd Plan & Research
Brainstorming, planning, and codebase exploration
/ck:brainstorm Brainstorm with problem-first inversion + scout-first gates
Mandatory codebase scout before any question; inverts a preselected solution to name the underlying problem first. Brutal-honest tradeoff analysis. --html writes an editorial-magazine report next to the markdown; --wiki publishes outputs to AgentWiki.
--html --wiki /ck:advise beta Honest advisory interview before planning or implementation
Takes a raw idea, problem, or URL (GitHub issue/PR, spec, doc) and interrogates it until real requirements and goals surface, then returns unfiltered advice. Advisory only — does not implement code or execute the advice. Output modes: --html, --md, --wiki, --github; --agent for autonomous single-run counsel.
[prompt-or-url] --html --md --wiki --github --agent /ck:research Technical research and technology evaluation
Best practices, solution design.
[topic] /ck:scout Quick codebase scan
Parallel agents for file discovery. Use ext flag to filter by file type.
[search-target] [ext] /ck:plan Create detailed plans with HTML, GitHub, and AgentWiki output
Modes --hard / --deep / --parallel / --two auto-run red-team; --deep also auto-runs validate. --html writes a self-contained interactive plan.html; --github creates or updates a GitHub issue with branch, summary, and acceptance criteria; --wiki publishes the plan to AgentWiki. Subcommands: archive, red-team, validate.
archive red-team validate --auto --fast --hard --parallel --two --no-tasks --deep --tdd --html --github --wiki /ck:sequential-thinking Step-by-step reasoning with self-correction
Multi-step analysis, hypothesis verification, adaptive planning.
/ck:problem-solving Break through deadlocks on complex problems
Break complex loops, test assumptions, simplify the problem.
/ck:loop Autonomous optimization loop
Run N iterations against a metric, learn from git history, auto-keep/discard changes. For coverage, performance, bundle size.
/ck:autoresearch router Autoresearch family router — learn the pattern, pick the right skill
Meta-framework for autonomous goal-directed iteration. Routes to 4 specialized skills: ck-loop, ck-predict, ck-scenario, ck-security.
/ck:predict 5 expert personas debate proposed changes before implementation
Catches architectural, security, performance, and UX issues early. Chain modes: --chain reason (subjective refinement), --chain probe (constraint discovery).
--files --chain reason --chain probe /ck:xia Safely study, extract, and port features from other repositories
Analyze reference implementations (GitHub/Local) to understand feature architecture. Integrates a Challenge Framework to evaluate risks before implementation, preventing blind copy-pasting.
<github-url|owner/repo|local-path> [feature] --compare --copy --improve --port --auto --fast /ck:issue-to-plan beta GitHub issue → audited plan (planning only)
Scouts the codebase, runs a hard brainstorm/audit gate, then only if it passes: plan generation, validate, red-team, pushed plan branch, and issue handoff. Stops before implementation — no cook, ship, or PR. Flags: --repo, --plan-ready-label, --decision-label.
<github-issue-url | issue-number> --repo owner/name --plan-ready-label --decision-label /ck:goal-warmup beta Preflight long-running goal work before execution
Locks an Outcome Contract, surfaces blockers, and reviews scope drift. Ends in exactly one of Ready, Blocked, or Decision required. Never auto-starts /goal. Use --fast for a lighter preflight.
"<goal>" --fast /ck:fable-thinking beta Evidence-grounded reasoning protocol for high-stakes analysis
Claude Fable 5 reasoning discipline as an executable protocol: multi-hypothesis diagnosis, concrete simulation, adversarial self-review, and calibrated claims. Use for debugging, reviews, architecture decisions, contested claims, constrained writing, and high-stakes analysis.
[task or question to reason through] /ck:research-prompt beta Self-contained research brief for human or AI researcher
Writes one self-contained paragraph that gives a researcher enough context for decision-ready, source-backed findings without follow-up. Writes the brief only — does not perform the research.
<research topic | decision> Design & Frontend
UI/UX design, frontend development, 3D graphics, and visual tools
/ck:frontend-design Create complete frontend interfaces
From designs, screenshots, or videos.
/ck:frontend-development Build React/TypeScript frontend
Modern patterns, MUI v7, TanStack.
/ck:ui-styling Style UI with shadcn/ui
Radix UI + Tailwind CSS. Dark mode, design system.
/ckm:design Brand identity, logo, CIP, slides, banner, icon
55 logo styles, 50 CIP deliverables, social media images. Generate images with Gemini AI.
/ck:stitch AI design generation via Google Stitch
Generate high-fidelity UI from text prompts. Export Tailwind/HTML. Design-to-code pipeline.
generate export --format all --format html --device mobile /ck:ui-ux-pro-max 50+ styles, 161 color palettes, 57 font pairings
161 product types, 99 UX guidelines, 25 chart types, 10 stacks.
/ck:web-design-guidelines Evaluate UI against Web Interface Guidelines
Check code against design guideline standards.
/ck:threejs Build 3D web apps with Three.js
WebGL/WebGPU. 556 examples, 60 API classes.
/ck:shader Write GLSL fragment shaders
Procedural graphics and generative art. Works with Three.js.
/ck:mermaidjs-v11 Flowcharts, diagrams, Gantt, timeline
Sequence, class, ER, state diagrams using Mermaid v11.
/excalidraw Hand-drawn diagrams with Excalidraw
Architecture, flowcharts, system designs in hand-drawn style. Auto-diagram mode for zero-config codebase visualization.
AI Generation & Multimodal
Image/video generation, AI analysis, and media processing
/ck:ai-multimodal Analyze images/audio/video with Gemini API
Generate images (Imagen 4, Nano Banana 2), videos (Veo 3, Hailuo).
/ck:ai-artist Generate images via Nano Banana
129 curated prompts. Validation required.
--mode search --mode creative --mode wild --mode all --provider auto --provider google --provider openrouter --skip /ck:media-processing FFmpeg, ImageMagick, AI background removal
Video/audio processing, images, and RMBG operations.
/ck:remotion Create programmatic videos in React
Build video components with React and render to MP4.
/ck:html-video Local MP4 videos from HTML/CSS/JS templates
Renders HTML templates to real MP4 via headless Chromium and ffmpeg (nexu-io/html-video). Covers source checkout setup, template discovery, studio customization, preview, and render verification.
[video brief, source URL, repo, template id, or project id] /ck:document-skills beta Route Office document work to the right CK skill
Router for Word, PDF, PowerPoint, and spreadsheet tasks. Dispatches to /ck:docx, /ck:pdf, /ck:pptx, or /ck:xlsx when the document type is unclear or the user asks for Office processing generally.
[document task] Build & Ship
Backend, frameworks, databases, DevOps, mobile, and deployment
/ck:bootstrap Full project initialization from scratch
Modes: full (default interactive), auto (explicit autonomous), fast (skip research), parallel (multi-agent).
--full --auto --fast --parallel /ck:backend-development Node.js, Python, Go APIs
NestJS, FastAPI, Django. REST/GraphQL/gRPC.
/ck:web-frameworks Next.js monorepo, Turborepo
Supports App Router, RSC, SSR, ISR.
/ck:databases MongoDB, PostgreSQL, SQL/NoSQL
Queries, aggregation, indexes, migrations.
/ck:tanstack TanStack Start, Form, AI
Streaming/chat with TanStack AI.
/ck:react-best-practices Optimize React/Next.js performance
From Vercel Engineering patterns.
/ck:devops Cloudflare, Docker, GCP, Kubernetes
Deploy Workers, R2, D1, Cloud Run.
/ck:ship Ship feature branches with automated pipeline
merge target -> test -> review -> journal (optional AgentWiki publish) -> push -> PR. Use for official main/master or beta dev/beta releases.
official beta --skip-tests --skip-review --skip-journal --skip-docs --dry-run /ck:mobile-development React Native, Flutter, Swift, Kotlin
Supports SwiftUI and Jetpack Compose.
/ck:team Orchestrate Agent Teams
Multi-session parallel collaboration. Templates: research, implement, review, debug.
<template> <context> --devs N --researchers N --reviewers N --delegate /ck:deploy Auto-detect and deploy to 15 platforms
Vercel, Netlify, Cloudflare, Railway, Fly.io, Render, Heroku, GitHub Pages, AWS, GCP, and more.
[platform] [environment] /ck:cook All-in-one feature implementation with hard gates
Scout-first → exact-requirements → plan → implement → test → no-side-effects verify → review. Gates block unsafe completion. --no-test override available.
--interactive --fast --parallel --auto --no-test --tdd /ck:vibe Full autonomous product-development pipeline
Run a complete pipeline from GitHub issue or feature request through worktree, TDD plan, cook/fix, code review, ship PR, review-pr, and optional merge with CI convergence.
<github-issue-url | feature request> --ship --beta Git & Version Control
Git operations, conventional commits, worktrees, and branch management
/ck:git Git operations with conventional commits + PR merge automation
Auto-splits by type/scope. Security scan before commit. merge-pr runs review-pr --fix --reply on each PR, labels ready to ship once merge-ready, merges, and watches CI to green.
cm cp pr merge merge-pr /ck:worktree Isolated git worktrees for parallel development
Create feature branches in separate directories. Ideal for monorepos.
Test & Debug
Testing, code review, debugging, and edge-case scenarios
/ck:test Run local and UI tests
Run full suite, specific module, or visual UI tests on a live website.
[context] ui [url] /ck:code-review Evidence-based code quality review
Reviews pending diffs, PR #, commit hash, or full codebase scan. Spec → Quality → Verify protocol focuses on bugs, regressions, maintainability, reliability, and verification gaps.
[context] #PR COMMIT --pending codebase codebase parallel /ck:review-pr Review GitHub PRs with auto-fix and formal reply
Analyzes a PR for duplicate prior work, project standards, strategic necessity, correctness, security, breaking changes, code quality, and AI-slop patterns. --fix auto-remediates findings, resolves conflicts, and watches CI to green; --reply posts the review to GitHub via gh.
<PR number or URL> --fix --reply /ck:fix Structured diagnosis & fix with side-effect sweep
Scout → exact-root-cause (6 items) → assess → fix → verify → side-effect sweep across blast radius → prevent. Evidence-based RCA, never silent patch.
--auto --review --quick --parallel /ck:debug Systematic root cause analysis
Trace call stack, multi-layer validation. Use when you don't know where the issue is.
/ck:web-testing Testing with Playwright, Vitest, k6
E2E/unit/integration/load/security/visual/a11y tests.
/ck:agent-browser AI-optimized browser automation CLI
Efficient context snapshots for complex web interactions.
/ck:chrome-profile Browser automation on the user's real Chrome profile
Targets a real Google Chrome profile via Chrome DevTools MCP. CLI chrome-profile with profile discovery, live DevTools probing, setup playbooks, and URL-anchor tab selection. open --json returns a bind_selector for exact tab binding; --no-activate keeps focus on the current app; ambiguous profile keys fail closed. Use when automation needs real cookies, Google account, or a logged-in session.
open --json --no-activate /ck:scenario Generate edge cases across 12 dimensions with saturation mode
Pre-implementation risk discovery and QA planning. --saturation loops until novelty dries up — confirms exhaustive coverage.
--saturation --iterations Security & Intelligence
Security audits, vulnerability scanning, authentication, and threat intelligence
/ck:security STRIDE + OWASP audit with red-team persona discovery loop
4 attacker personas (security adversary, supply chain, insider, infrastructure). --red-team for persona discovery loop. --fix for auto-remediation.
--fix --red-team --iterations /ck:security-scan Scan codebase for security vulnerabilities
Automated security analysis with remediation suggestions.
[scope] --secrets-only --deps-only --full /ck:better-auth TypeScript auth: OAuth, 2FA, passkeys
Supports RBAC, WebAuthn, MFA.
/ck:cti-expert CTI/OSINT investigation toolkit
Cyber threat intelligence and open-source investigation: case timelines, recon, breach checks, digital footprint, image/blockchain forensics, and GitHub OSINT via /github-osint. No API keys required.
/github-osint [target] --yolo --case --sweep --query --flow Docs & Content
Documentation management, content writing, and file viewing
/ck:docs Analyze codebase and manage documentation
Initialize, update, or summarize project documentation.
init update summarize /ck:docs-seeker Search library/framework documentation
Via llms.txt (context7.com). Fastest way to get up-to-date docs.
/ck:copywriting Headlines, CTAs, landing pages
Email copy, writing styles, conversion-focused copy.
[copy-type] [context] /ck:mintlify Documentation site with Mintlify
MDX components, themes, OpenAPI.
/ck:show-off Generate stunning HTML showcases for demos and social media
Multi-section scrollable pages with parallax, theme toggle, bilingual (VI/EN). Auto-captures sections as images. Persists workflow preferences (screenshots, publishing, languages) at ~/.claude/show-off/preferences.json across sessions.
/ck:preview View files or create visual explanations
Markdown or self-contained HTML. Modes: --explain, --slides, --diagram, --diff, --plan-review, --recap. Add --html for browser output.
--explain --slides --diagram --ascii --html --diff --plan-review --recap /ck:tech-graph Generate publication-quality SVG technical diagrams
8 design styles (modern, minimal, neon, retro, Dark Luxury premium dark editorial, etc.) and 10 templates (architecture, sequence, ER, flow). SVG output ready for docs, slides, blog posts.
/ck:markdown-novel-viewer Book-style reading via HTTP server
Comfortable reading for markdown files.
/ck:llms Generate llms.txt from docs or codebase
Follows llmstxt.org spec. Includes reference files and generation script.
--full --output /ck:interview-docs beta Interview-driven durable project documents
Turns the user's own knowledge, taste, and decisions into maintained docs (README, ADR, strategy, principles). User answers become the content — does not invent from code inspection or AI proposals.
<vision | document-path | topic> MCP & Integration
MCP servers, payment, Shopify, and AI agent integrations
/ck:mcp-builder Build MCP servers
FastMCP (Python), MCP SDK (Node/TypeScript).
/ck:use-mcp Discover and execute MCP tools
Two paths: Gemini CLI (LLM-driven) or direct scripts (deterministic). Replaces mcp-management.
/ck:payment-integration SePay, Polar, Stripe, Paddle, Creem.io
VietQR and international payment providers.
/ck:shopify Shopify apps and Liquid templates
Polaris UI, checkout customization.
/ck:google-adk-python Build AI agents with Google ADK Python
A2A protocol, MCP tools integration.
/ck:agentize Convert codebase into AI agent-friendly CLI + MCP server
Wrap code as CLI/MCP with shared core/. Agent-centric design. Modes: --both, --mcp, --cli.
--both --mcp --cli --auto --ask Session Management
Session review, journaling, and context optimization
/ck:watzup Session hand-off summaries from recent changes
What shipped, what's in flight, what's next. Use at EOD or before context handoff.
/ck:journal Write change analysis journal
Record thoughts and technical decisions.
/ck:context-engineering Monitor context usage and optimize tokens
Check limits and optimize context window consumption.
/ck:retro Data-driven sprint retrospective
Git metrics, health indicators, and actionable recommendations.
[timeframe] --compare --team --format html|md /ck:handoff beta Compact redacted handoff for a fresh agent session
Creates a factual, compact conversation handoff so a new agent can continue with minimal rediscovery. Preserves decisions and blockers — not a command list. Use when switching context or ending a work session.
[next-session focus] Project Tracking
Progress tracking, kanban boards, and plan management
/ck:project-management Track progress and manage Claude Tasks
Generate reports, coordinate documentation. Tasks: status, hydrate, sync, report.
status hydrate sync report /ck:plans-kanban Planning board with progress tracking
Visualize timeline and progress.
/ck:project-organization Organize project structure and file layout
Analyze and restructure project directories for maintainability.
Utilities
Skill management, code packaging, semantic analysis, and tooling
/ck:skill-creator Create/update Claude skills
3 sub-agents (analyzer, comparator, grader). Eval infrastructure with benchmarks.
/ck:find-skills Discover and install skills from the ecosystem
Search by capability or task description.
/ck:repomix Package repositories into AI-friendly files
Output in XML, markdown, plain, or JSON format.
--style xml --style markdown --style plain --style json /ck:gkg Semantic code analysis
GitLab Knowledge Graph integration.
/ck:graphify Build queryable knowledge graphs
Turn code, docs, papers, or images into knowledge graphs. Tree-sitter AST for 20 languages.
[path] --mcp --report --watch /ck:ghpm GitHub project management for humans and AI agents
Use GitHub as single source of truth for project work — issue/task planning, Projects boards, handoff status, gh/API automation, and CI-driven tracking. Modes: bootstrap, intake, execute, handoff, audit.
/ck:coding-agent-orchestration beta Coordinate multiple coding agents across one workflow
Plans the human workflow for choosing Claude Code, Codex, Gemini, OpenCode, Cursor, Amp, Droid, or Antigravity — splitting work safely, defining handoffs, and consolidating plans, reviews, tests, and PRs behind one integrator. Modes: --single, --sequential, --parallel, --review-loop.
[task or workflow] --single --sequential --parallel --review-loop /ck:codex-goal beta Guide Codex /goal with a verifiable stop condition
Codex-native /goal guidance for durable objectives with a clear stopping condition and validation loop. Not a safety boundary, product-decision substitute, or unbounded backlog runner.
<objective | goal draft> /ck:folder-context beta Subfolder CLAUDE.md + AGENTS.md for local conventions
Creates a compact subfolder CLAUDE.md and linked AGENTS.md for durable local conventions. Never edits root CLAUDE.md or AGENTS.md. Use when a folder needs agent context beyond the project root.
<target-folder> Related guide
Finding Your Unknowns
Not sure which command category you need? Finding Your Unknowns maps each kind of unknown to the right ClaudeKit command before you browse the full catalog.