Workflow Recipes

Ready-to-use command sequences for common development tasks.

New to ClaudeKit?
These workflows consist of sequences of commands that work together to accomplish common tasks. Perfect for getting started!

Workflow Recipes

Build a New Feature

Beginner

Best for: Adding new functionality to your app

~15-30 min 4 steps
1
/ck:brainstorm skill Explore ideas (skill)

Use /ck:brainstorm or say "brainstorm ideas for [feature]" to explore possibilities

2
/ck:plan Create implementation plan

AI creates a detailed step-by-step plan for building your feature

3
Review → /clear → Implement

Review plan, run /clear to free context (mandatory), then implement

4
/ck:cook @plan.md skill Implement the plan (skill)

AI writes the code following the plan, runs tests, and reviews the work

/clear is mandatory after /ck:plan before /ck:cook

Fix a Bug

Beginner

Best for: Fixing errors and unexpected behavior

~5-15 min 3 steps
1
/ck:debug skill Investigate the issue (skill)

AI analyzes your code to find the root cause of the problem

2
/ck:fix skill Apply the fix (skill)

AI intelligently routes to specialized fix and applies the solution

3
/ck:test Verify the fix

Run tests to make sure the bug is fixed and nothing else broke

/ck:fix Flags
--auto

Auto-apply fix without confirmation

--review

Review fix before applying

--quick

Fast fix without deep analysis

--parallel

Fix multiple issues in parallel

/ck:debug & /ck:fix: intelligent routing with flags --auto, --review, --quick, --parallel

Quick Implementation

Intermediate

Best for: Small features when you know what you want

~10-20 min 1 steps
1
/ck:cook "your task" skill All-in-one skill

AI researches, plans, implements, tests, and reviews the feature automatically

What /ck:cook "your task" does automatically:
  • Research best approaches and technologies
  • Create and execute implementation plan
  • Write and test the code
  • Review for quality and best practices
/ck:cook Flags
--interactive

Step-by-step with approval (default)

--fast

Skip research, quick implementation

--parallel

Run phases in parallel

--auto

Auto-approve all steps

--no-test

Skip test runs after coding

--interactive is default, uses native Claude Tasks API.

Start New Project

Advanced

Best for: Creating a complete app from scratch

~1-2 hours 1 steps
1
/ck:bootstrap "describe your app" Complete project setup

AI builds your entire project: research, architecture, design, implementation, and documentation

What /ck:bootstrap "describe your app" does automatically:
  • Research and tech stack selection
  • Project structure and architecture
  • UI/UX design and wireframes
  • Complete implementation with tests
  • Comprehensive documentation
Warning: This uses significant AI tokens

Create Video Content

Intermediate

Best for: Programmatic video creation with React

~20-40 min 2 steps
1
/ck:remotion skill Video creation (skill)

Create videos programmatically with React and Remotion

2
render Render output

Export to MP4, GIF, or image sequences

What /ck:remotion does automatically:
  • Animations and transitions
  • Text animations and captions
  • 3D graphics integration
  • Audio synchronization
Creates React-based video compositions

Visual Documentation

Beginner

Best for: Creating visual explanations and diagrams for your plan

~10-20 min 3 steps
1
/ck:plan Create plan first

Create a structured plan. Visuals are saved into the plan directory

2
/ck:preview --explain "topic" skill Generate explanation (skill)

Create ASCII + Mermaid diagrams with prose explanation for your topic

3
/ck:preview --diagram "topic" skill Generate focused diagram (skill)

Create a focused Mermaid + ASCII diagram for a specific data flow or architecture

What /ck:preview does automatically:
  • ASCII + Mermaid diagrams (--explain)
  • Presentation format (--slides)
  • Focused diagrams (--diagram)
  • Terminal-friendly output (--ascii)
/ck:preview generates visual content. Also: --slides, --ascii

Code Review with Edge Cases

Intermediate

Best for: Thorough code review with edge case scouting

~20-30 min 4 steps
1
/ck:cook @plan.md skill Implement the plan (skill)

AI writes the code following the plan with auto test & review cycles

2
/ck:scout skill Scout edge cases (skill)

AI scouts affected files, data flows, error paths, and boundary conditions

3
Code-reviewer review

Code-reviewer subagent reviews findings from scout and evaluates code quality

4
/ck:git cm Merge & commit

Commit the reviewed code with conventional commit message

What /ck:scout does automatically:
  • Edge case detection via /ck:scout
  • Boundary condition analysis
  • Data flow & error path scouting
  • Automated code-reviewer integration
/ck:scout integrates with code-reviewer for edge case detection before review

Plan + Validate + Implement

Intermediate

Best for: Validated plans with auto-propagated decisions

~20-40 min 4 steps
1
/ck:plan Create plan

AI creates a detailed implementation plan with phases

2
/ck:plan validate Validate plan decisions

Interview-style validation gate. Decisions auto-propagate to phase files

3
/clear (mandatory)

Free context before implementation. Mandatory step

4
/ck:cook @plan.md skill Implement validated plan (skill)

AI implements with validated decisions already propagated to each phase

What /ck:plan validate does automatically:
  • Interview-style plan validation
  • Auto-propagation to phase files
  • Validated decisions guide implementation
  • Reduced rework from unclear plan decisions
/ck:plan Flags
--hard

Complex multi-phase plan with red-team review

--parallel

Plan designed for parallel agent execution

--two

Two-phase plan (plan → implement)

validate

Interview-style validation gate

red-team

Spawn adversarial reviewers

/ck:plan validate decisions auto-propagate to phase files

Agent Teams (Parallel)

Advanced

Best for: Large tasks with multiple parallel agents

~30-60 min 2 steps
1
/ck:plan --hard "feature" Create plan with phases

Create a detailed plan with parallelizable phases for team execution

2
/ck:team cook @plan skill Parallel team execution (skill)

Spawn parallel dev agents, each handling a phase. Auto test → review → merge

What /ck:plan --hard "feature" does automatically:
  • Parallel dev agents (--devs N)
  • Auto test → review → merge pipeline
  • Event-driven hooks + agent memory
  • Also: /ck:team research, /ck:team review, /ck:team debug
/ck:team Flags
--devs N

Number of parallel dev agents (default: 2)

--tester

Include dedicated tester agent

--reviewer

Include code reviewer agent

--worktree

Run each agent in isolated git worktree

Requires CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1 in settings.json

Adversarial Plan Review

Advanced

Best for: Finding flaws in plans before implementation

~10-20 min 2 steps
1
/ck:plan --hard "feature" Create plan

Create a detailed plan. Hard/parallel/two modes auto-run red-team after creation

2
/ck:plan red-team plans/ skill Adversarial review (skill)

Spawn hostile reviewers: Security, Failure Mode, Assumption Destroyer, Scope Critic

What /ck:plan --hard "feature" does automatically:
  • Security Adversary (auth bypass, injection, OWASP)
  • Failure Mode Analyst (race conditions, data loss)
  • Assumption Destroyer (unstated deps, false claims)
  • Scope & Complexity Critic (over-engineering, YAGNI)
Auto-scales reviewers based on plan complexity (2-4 adversarial lenses)

Research & Documentation

Beginner

Best for: Research topics and create technical documentation

~10-20 min 3 steps
1
/ck:research "topic" skill Deep research (skill)

AI researches the topic thoroughly using web search and documentation

2
/ck:docs-seeker "library" skill Search library docs (skill)

Search official documentation via llms.txt for up-to-date API info

3
/ck:docs skill Generate project docs

Create or update project documentation based on codebase analysis

What /ck:research does automatically:
  • Web search and synthesis
  • Library documentation lookup
  • Project documentation generation
  • Technical writing assistance
/ck:docs-seeker uses context7 for latest library documentation

Security Audit

Intermediate

Best for: Finding security vulnerabilities and secrets

~15-25 min 3 steps
1
/ck:security-scan skill BETA Scan for vulnerabilities (skill)

Scan codebase for OWASP issues, hardcoded secrets, and dependency vulnerabilities

2
/ck:code-review --security skill Security-focused review

Deep code review focusing on authentication, authorization, and data handling

3
/ck:fix --security skill Apply security fixes

AI applies recommended security fixes with detailed explanations

What /ck:security-scan does automatically:
  • Hardcoded secrets detection
  • Dependency vulnerability scan
  • OWASP Top 10 coverage
  • Security fix recommendations
Detects SQL injection, XSS, CSRF, and other OWASP Top 10 issues

Database Operations

Intermediate

Best for: Database schema design and migrations

~15-30 min 3 steps
1
/ck:databases "schema design" skill Design schema (skill)

Design database schema with relationships, indexes, and constraints

2
/ck:plan "migration" Plan migration

Create a safe migration plan with rollback strategy

3
/ck:cook @plan.md skill Execute migration

Implement migration with proper error handling and validation

What /ck:databases does automatically:
  • Schema design with relationships
  • Index optimization
  • Migration script generation
  • Query performance analysis
Supports MongoDB, PostgreSQL, MySQL, and SQLite

DevOps & Deployment

Advanced

Best for: Setting up CI/CD and deployment pipelines

~20-40 min 3 steps
1
/ck:devops "setup CI/CD" skill Configure DevOps (skill)

Set up CI/CD pipelines for GitHub Actions, GitLab CI, or other platforms

2
/ck:deploy skill BETA Deploy to platform (skill)

Deploy to Cloudflare, Vercel, GCP, or Kubernetes with auto-detection

3
/ck:test --e2e Run E2E tests

Verify deployment with end-to-end tests

What /ck:devops does automatically:
  • CI/CD pipeline generation
  • Multi-platform deployment
  • Environment configuration
  • Rollback and monitoring setup
Supports Docker, Kubernetes, serverless, and container deployments

Pro Tips

Save tokens with /clear
Use /clear to clear conversation history before starting implementation
Reference files with @
Use @plan.md to reference specific files in commands
Use /ck:cook for implementation
After planning with /ck:plan, run /clear then /ck:cook to implement your plan
Scout before coding
Use /ck:scout to quickly find relevant files in your codebase
Debug before fixing
Use /ck:debug first to find root cause, then /ck:fix when ready
Discover commands with /ck:ck-help
Type /ck:ck-help to see all commands & skills organized by category

Need more workflows? See full docs from official ClaudeKit site.