Skip to main content

Skills Reference

SpecWeave provides 100+ specialized skills that activate automatically based on context or can be invoked directly. Skills are more powerful than commands -- they provide domain expertise, best practices, and guided workflows.

Three Ways to Invoke Any Skill
  1. Natural language -- just describe what you want (e.g., "design the system architecture")
  2. Slash command -- sw:skill-name in Claude Code
  3. Keyword -- skill-name without prefix in Cursor, Copilot, and other AI tools

How Skills Work​

Invocation methods:

  1. Natural language (easiest): Just describe what you need -- skills activate based on keywords
  2. Slash command (precise): sw:skill-name or sw-plugin:skill-name in Claude Code
  3. CLI keyword (cross-tool): skill-name without prefix in Cursor, Copilot, and other AI tools
  4. Chaining: Skills can invoke other skills for complex workflows

Core Skills (Planning & Management)​

These skills form the foundation of spec-driven development.

Natural LanguageClaude CodeOther AIDescription
"let's build X", "add feature X"sw:incrementincrementPlan and create increments
"write specs", "define requirements"sw:pmpmProduct Manager for specs
"design the system", "architecture"sw:architectarchitectSystem architect for designs
"coordinate all agents"sw:role-orchestratorrole-orchestratorMulti-agent coordination
"plan the roadmap"sw:roadmap-plannerroadmap-plannerProduct roadmap planning
"regenerate the plan"sw:planplanGenerate plan.md and tasks.md
"brainstorm X", "explore ideas"sw:brainstormbrainstormMulti-perspective ideation

brainstorm​

Purpose: Explore a problem from multiple angles before committing to an implementation path. Uses Tree of Thought divergent exploration with selectable cognitive lenses.

"Brainstorm approaches for real-time notifications"
# Options
sw:brainstorm "auth system" --depth deep              # All 5 phases + deepening
sw:brainstorm "payment gateway" --lens six-hats       # Specific lens
sw:brainstorm "API design" --depth quick              # Fast 3-approach comparison

5-Phase Flow:

  1. Frame — Problem statement, starbursting (5W1H), clarifying questions
  2. Diverge — Generate 4-6 approaches via selected cognitive lens
  3. Evaluate — Comparison matrix with scoring and explicit recommendation
  4. Deepen (deep only) — Abstraction laddering, analogies, pre-mortem analysis
  5. Output — Save persistent brainstorm doc, offer handoff to sw:increment

Depth Modes:

  • quick — 3 inline approaches, skip lenses (Phase 1+3)
  • standard — Default lens, 4-6 approaches (Phase 1-3+5)
  • deep — Multiple lenses via parallel subagents, all 5 phases

Cognitive Lenses:

  • default — Independent parallel generation (conservative, bold, speed, extensibility)
  • six-hats — Six Thinking Hats (facts, feelings, caution, optimism, creativity, process)
  • scamper — SCAMPER transformations (substitute, combine, adapt, modify, repurpose, eliminate, reverse)
  • triz — Constraint inversion (negate core assumptions)
  • adjacent — Adjacent Possible (what recently became feasible)

Additional features:

  • --resume — Pick up a previous brainstorm session
  • --criteria — Custom evaluation criteria (e.g., brand-fit,audience-reach,cost)
  • TRIZ lens includes 13 inventive principles adapted for software
  • Adjacent Possible lens uses live web search to ground ideas in reality

Output: Persistent brainstorm document at .specweave/docs/brainstorms/YYYY-MM-DD-topic.md with idea tree for revisiting later.

Full guide: Brainstorming with Cognitive Lenses

increment​

Purpose: Plan and create SpecWeave increments with PM and Architect collaboration.

"Let's build user authentication with JWT"

What it does:

  • Detects tech stack automatically
  • Runs PM-led spec creation
  • Creates plan.md with architecture
  • Generates tasks.md with proper AC-IDs

pm​

Purpose: Product Manager for spec-driven development.

sw:pm  # Activate PM guidance

Covers:

  • User story writing with proper format
  • Acceptance criteria with testable AC-IDs
  • MVP prioritization (MoSCoW, RICE)
  • Requirements refinement

architect​

Purpose: System architect for scalable, maintainable designs.

sw:architect  # Get architecture guidance

Covers:

  • ADR (Architecture Decision Record) writing
  • Microservices vs monolith decisions
  • Database schema design
  • Technology selection trade-offs

role-orchestrator​

Purpose: Multi-agent orchestration for complex features.

sw:role-orchestrator "Build SaaS dashboard with payments"

Coordinates:

  • PM for requirements
  • Architect for design
  • DevOps for infrastructure
  • QA for testing strategy
  • Security for compliance

roadmap-planner​

Purpose: Product roadmap and feature prioritization.

sw:roadmap-planner  # Roadmap planning guidance

Covers:

  • Quarterly and annual roadmap planning
  • Feature prioritization (RICE, MoSCoW)
  • Dependency mapping across increments
  • Milestone tracking

plan​

Purpose: Generate plan.md and tasks.md for an existing increment.

sw:plan 0007  # Regenerate plan and tasks for increment 0007

Covers:

  • Generating plan.md with architecture decisions
  • Creating tasks.md with proper AC-ID linking
  • Rebuilding task breakdown from spec

Execution & Workflow Skills​

Control the development workflow from start to finish.

Natural LanguageClaude CodeOther AIDescription
"start implementing", "execute tasks"sw:dodoExecute tasks manually
"ship while I sleep", "autonomous mode"sw:autoautoAutonomous execution
"parallel agents"sw:auto-parallelauto-parallelMulti-agent parallel execution
"check auto progress"sw:auto-statusauto-statusCheck autonomous progress
"stop auto mode"sw:cancel-autocancel-autoEmergency stop
"what's the status?", "show progress"sw:progressprogressShow increment status
"what's next?"sw:nextnextSmart workflow transition
"we're done", "close it"sw:donedoneClose increment
"check quality", "validate it"sw:validatevalidateRule-based validation
"quality check"sw:qaqaAI quality gate
"review my work", "critique the code"sw:grillgrillImplementation auditor

auto​

Purpose: Autonomous execution that runs for hours.

"Ship it while I sleep"

Features:

  • Continuous loop: Read task → Implement → Test → Fix → Next
  • Self-healing: Retries failures up to 3 times
  • Living docs sync: Updates documentation automatically
  • Proven for 2-3 hour sessions

grill​

Purpose: Comprehensive implementation audit using parallel subagents.

"Audit the code quality"
sw:grill src/auth           # Specific module
sw:grill --focus security   # Focus on security

Audits:

  • Structure and organization
  • Code quality and patterns
  • Consistency across codebase
  • Documentation completeness
  • Dependency health
  • Test coverage
  • Security vulnerabilities

Quality & Testing Skills​

Ensure code quality and test coverage.

Natural LanguageClaude CodeOther AIDescription
"write failing tests first"sw:tdd-redtdd-redWrite failing tests first
"make the tests pass"sw:tdd-greentdd-greenMinimal implementation to pass
"refactor the code"sw:tdd-refactortdd-refactorImprove code quality
"TDD", "test-driven development"sw:tdd-cycletdd-cycleFull TDD workflow
"review code", "code review"sw:code-reviewercode-reviewerElite multi-agent code review
"parallel agents", "team work"sw:team-leadteam-leadPhase-agnostic orchestrator
"judge the quality"sw:judge-llmjudge-llmLLM-as-Judge quality assessment
"debug this"sw:debugdebugSystematic 4-phase debugging
"e2e tests", "playwright tests"sw:e2ee2eEnd-to-end testing

code-reviewer​

Purpose: Elite multi-agent code review. Spawns up to 6 specialized reviewer agents in parallel, then aggregates findings into a unified report with deduplication and severity ranking.

"Review my code changes"
sw:code-reviewer --pr 42            # Review a specific PR
sw:code-reviewer --changes          # Review uncommitted changes
sw:code-reviewer --increment 0042   # Review changes from an increment
sw:code-reviewer --cross-repo       # Aggregate across umbrella repos

Specialized Reviewers (spawned in parallel):

  • Logic — bugs, edge cases, error handling, race conditions
  • Security — OWASP Top 10, auth, secrets, injection
  • Performance — N+1 queries, memory leaks, blocking ops
  • Silent Failures — empty catches, swallowed errors, missing .catch()
  • Type Design — unsafe assertions, overly broad types, missing invariants
  • Spec Compliance — AC verification, scope creep detection (increment scope only)

Smart routing: Not all 6 reviewers run every time. Reviewers are selected based on what files changed (e.g., TypeScript files trigger the type reviewer, database files trigger performance).

vs sw:grill: Grill is increment-scoped and runs during closure (mandatory gate). Code-reviewer is general-purpose — use it anytime on any scope.

team-lead​

Purpose: Phase-agnostic orchestrator for parallel multi-agent work. Auto-detects operating mode from your intent.

"Build this with parallel agents"
sw:team-lead "Brainstorm auth approaches"       # Brainstorm mode
sw:team-lead "Plan the payment system"          # Planning mode
sw:team-lead "Review recent changes"            # Review mode
sw:team-lead "Research caching strategies"      # Research mode
sw:team-lead --mode plan "user dashboard"       # Explicit mode override

6 Operating Modes:

ModeWhat It DoesIncrement?
BrainstormSpawns advocate + critic + pragmatist agents, synthesizes decision matrixNo
PlanningPM + Architect agents in parallel for richer specsCreates one
ImplementationDomain agents (frontend, backend, database) with contract-first spawningRequired
ReviewDelegates to sw:code-reviewer for parallel multi-agent reviewOptional
Research1-3 researcher agents exploring different facets of a topicNo
TestingTesting agents split by layer (unit, integration, E2E)Required

Frontend Skills​

Build modern web applications with best practices. These are available as community skills via verified-skill.com.

SkillSourceDescription
frontendcommunityGeneral frontend development
frontend-architectcommunityReact/Vue/Angular architecture
frontend-designcommunityUI/UX implementation
design-system-architectcommunityComponent library design
nextjscommunityNext.js App Router patterns
code-explorercommunityCodebase navigation

frontend-architect​

Purpose: Frontend architecture for React, Vue, or Angular.

sw:architect  # Frontend architecture guidance

Covers:

  • Component composition patterns
  • State management (Redux, Zustand, Pinia)
  • Performance optimization
  • Accessibility (a11y) requirements

nextjs​

Purpose: Next.js 14+ App Router patterns.

sw:architect  # Next.js guidance

Covers:

  • Server Components vs Client Components
  • Route handlers and middleware
  • Data fetching patterns
  • Image and font optimization

Backend Skills​

Build scalable APIs and services. These are available as community skills via verified-skill.com.

SkillSourceDescription
nodejs-backendcommunityNode.js/Express/Fastify APIs
python-backendcommunityPython/FastAPI/Django services
dotnet-backendcommunity.NET Core APIs
database-optimizercommunitySQL and NoSQL optimization

nodejs-backend​

Purpose: Node.js backend development.

/backend:nodejs  # Node.js guidance

Covers:

  • Express/Fastify patterns
  • Middleware architecture
  • Error handling
  • Authentication (JWT, sessions)

database-optimizer​

Purpose: Database performance and design.

/backend:database-optimizer  # DB optimization

Covers:

  • Query optimization
  • Index strategies
  • Connection pooling
  • PostgreSQL, MySQL, MongoDB patterns

Infrastructure & DevOps Skills​

Deploy and operate at scale. These are available as community skills via verified-skill.com.

SkillSourceDescription
devopscommunityCI/CD pipelines and deployment
observabilitycommunityLogging, metrics, tracing
infrastructurecommunityTerraform and cloud IaC
k8s-manifest-generatorcommunityKubernetes manifests
k8s-security-policiescommunityPod security and RBAC
helm-chart-scaffoldingcommunityHelm chart creation
gitops-workflowcommunityArgoCD/Flux workflows

devops​

Purpose: CI/CD and deployment automation.

/infra:devops  # DevOps guidance

Covers:

  • GitHub Actions, GitLab CI
  • Docker containerization
  • Vercel, Cloudflare deployment
  • Blue-green, canary deployments

k8s-manifest-generator​

Purpose: Generate Kubernetes manifests.

/k8s:deployment-generate  # K8s manifests

Generates:

  • Deployments, Services, Ingress
  • ConfigMaps, Secrets
  • HPA, PDB for resilience
  • Network policies

External Sync Skills​

Integrate with GitHub, JIRA, and Azure DevOps. All sync skills are part of the unified sw plugin.

SkillPluginDescription
github-syncswBidirectional GitHub sync
github-multi-projectswMulti-repo GitHub coordination
github-issue-standardswIssue formatting standards
pr-reviewswPull request review
jira-syncswBidirectional JIRA sync
jira-mapperswField mapping configuration
jira-resource-validatorswValidate JIRA resources
ado-syncswAzure DevOps sync
ado-multi-projectswMulti-project ADO coordination
ado-mapperswADO field mapping
ado-resource-validatorswValidate ADO resources

github-sync​

Purpose: Two-way sync between SpecWeave and GitHub Issues.

sw:github-sync 0007  # Sync increment 0007 to GitHub

Maps:

  • Feature → GitHub Milestone
  • User Story → GitHub Issue
  • Task → Issue checkbox

jira-sync​

Purpose: Bidirectional JIRA integration.

sw:jira-sync 0007  # Sync to JIRA

Maps:

  • Feature → JIRA Epic
  • User Story → JIRA Story
  • Task → JIRA Sub-task

Data & Streaming Skills​

Build event-driven architectures. These are available as community skills via verified-skill.com.

SkillSourceDescription
kafka-architectcommunityKafka architecture design
kafka-opscommunityKafka operations and monitoring
kafka-streams-topologycommunityStream processing topologies
confluent-kafka-connectcommunityKafka Connect configuration
confluent-schema-registrycommunityAvro/Protobuf schemas
confluent-ksqldbcommunityksqlDB streaming queries

kafka-architect​

Purpose: Kafka event-driven architecture.

/kafka:architect  # Kafka architecture guidance

Covers:

  • Topic design and partitioning
  • Producer/consumer patterns
  • Exactly-once semantics
  • Schema evolution

ML & AI Skills​

Build machine learning systems.

SkillSourceDescription
ml-specialistcommunityML model development
ml-engineercommunityML pipeline engineering
mlops-engineercommunityMLOps and model deployment
data-scientistcommunityData analysis and experimentation
imageswAI image generation (Google Imagen 4 / Pollinations.ai)
videoswAI video generation (Google Veo 3.1 / Pollinations.ai)
remotionswProgrammatic video from React with Remotion

ml-specialist​

Purpose: Machine learning model development.

/ml:engineer  # ML guidance

Covers:

  • Model selection
  • Feature engineering
  • Training pipelines
  • Model evaluation metrics

image​

Purpose: Generate images using AI. Google Imagen 4 (with GEMINI_API_KEY) or Pollinations.ai (free fallback).

sw:image "hero image for SaaS landing page"

Generates:

  • Hero images
  • Icons and logos
  • Product mockups

video​

Purpose: Generate videos using AI. Google Veo 3.1 (with GEMINI_API_KEY) or Pollinations.ai (free fallback).

sw:video "product demo showing the dashboard in action"

remotion​

Purpose: Create programmatic videos with Remotion (React components rendered to MP4).

sw:remotion "animated product launch video with logo reveal"
  • Illustrations

Mobile Skills​

Build cross-platform mobile apps. Install via: npx vskill install --repo anton-abyzov/vskill --plugin mobile

SkillPluginDescription
appstoremobile (vskill)App Store Connect automation and submission

appstore​

Purpose: App Store Connect automation and submission management.

/mobile:appstore  # App store guidance

Covers:

  • App Store Connect submission workflow
  • Screenshot and metadata management
  • Build and version coordination
  • Release management

Security & Compliance Skills​

Build secure, compliant systems. These are available as community skills via verified-skill.com.

SkillSourceDescription
securitycommunitySecurity assessment and hardening
security-patternscommunityReal-time security pattern detection
compliance-architecturecommunitySOC 2, HIPAA, GDPR compliance
pci-compliancecommunityPCI-DSS for payments

security​

Purpose: Security vulnerability assessment.

sw:security  # Security guidance

Covers:

  • OWASP Top 10
  • Authentication/authorization flaws
  • Injection vulnerabilities
  • Security headers

compliance-architecture​

Purpose: Enterprise compliance for regulated industries.

sw:compliance-architecture  # Compliance guidance

Covers:

  • SOC 2 Type II controls
  • HIPAA for healthcare
  • GDPR for EU data
  • PCI-DSS for payments

Payments Skills​

Build payment systems. These are available as community skills via verified-skill.com.

SkillSourceDescription
paymentscommunityPayment integration patterns
billing-automationcommunitySubscription and billing
pci-compliancecommunityPCI-DSS compliance

payments​

Purpose: Payment gateway integration.

/payments:payment-core  # Payment guidance

Covers:

  • Stripe integration
  • Webhook handling
  • Idempotency patterns
  • Refund workflows

Documentation Skills​

Create and maintain documentation. These are available as community skills via verified-skill.com.

SkillSourceDescription
docs-writercommunityTechnical documentation
living-docscommunityLiving documentation management
living-docs-navigatorcommunityNavigate project docs
technical-writingcommunityTechnical writing best practices
docusauruscommunityDocusaurus site management

docs-writer​

Purpose: Technical documentation generation.

sw:docs-writer  # Generate documentation

Creates:

  • API documentation
  • README files
  • User guides
  • Developer onboarding

docusaurus​

Purpose: Docusaurus documentation site.

/docs:docusaurus  # Docusaurus guidance

Covers:

  • MDX components
  • Sidebar configuration
  • Versioning
  • Search integration

Cost Optimization Skills​

Optimize cloud spending. These are available as community skills via verified-skill.com.

SkillSourceDescription
cost-optimizationcommunityGeneral cost reduction
aws-cost-expertcommunityAWS cost analysis
cloud-pricingcommunityMulti-cloud pricing comparison

cost-optimization​

Purpose: Cloud cost reduction strategies.

/cost:cost-optimization  # Cost analysis

Covers:

  • Right-sizing instances
  • Reserved capacity planning
  • Spot instance strategies
  • Storage tier optimization

Utility Skills​

General-purpose development utilities. These are available as community skills via verified-skill.com.

SkillSourceDescription
code-simplifiercommunityCode clarity improvement
reflectcommunityLearning from corrections
translatorcommunityContent translation
context-loadercommunityContext efficiency
detectorcommunitySpecWeave workflow detection
lspcommunityCode navigation (find refs, go to def)
brownfield-analyzercommunityExisting project analysis
brownfield-onboardercommunityProject migration
frameworkcommunitySpecWeave framework guide
service-connectcommunityExternal service connection

code-simplifier​

Purpose: Improve code clarity without changing behavior.

sw:code-simplifier  # Simplify recent changes

Does:

  • Removes unnecessary complexity
  • Improves variable naming
  • Extracts functions appropriately
  • Never changes functionality

lsp​

Purpose: Code navigation (LSP workaround for Claude Code v2.1.0+).

sw:lsp refs MyFunction     # Find references
sw:lsp def MyClass         # Go to definition
sw:lsp hover file.ts 42 10 # Type info at position

Release & Versioning Skills​

Manage releases and versions.

SkillPluginDescription
release-expertswRelease orchestration and version alignment

release-expert​

Purpose: Full release orchestration, version alignment, and coordination.

sw:release-expert  # Release guidance

Covers:

  • Semantic versioning and version alignment
  • Multi-repo release coordination and waves
  • RC lifecycle management
  • Changelog generation and rollback procedures

Diagrams & Visualization Skills​

Create technical diagrams.

SkillPluginDescription
diagramsswArchitecture diagram design and generation

diagrams​

Purpose: Create and generate Mermaid/C4 diagrams.

sw:diagrams  # Create diagrams

Creates:

  • C4 architecture diagrams (Context, Container, Component)
  • Sequence diagrams
  • Entity relationship diagrams
  • Deployment diagrams

Workflow Automation Skills​

Automate with n8n. These are available as community skills via verified-skill.com.

SkillSourceDescription
n8n-kafka-workflowscommunityKafka + n8n automation

n8n-kafka-workflows​

Purpose: Event-driven workflow automation.

# Available as community skill — install from verified-skill.com

Creates:

  • Kafka trigger nodes
  • Event processing workflows
  • Webhook integrations

Installing Plugins​

The unified sw plugin ships with SpecWeave and installs automatically via specweave init. It includes all 44 core skills (GitHub, JIRA, ADO sync, release, diagrams, media, and more).

Additional domain plugins are available from the vskill marketplace:

# Install the SpecWeave core plugin (usually handled by specweave init)
npx vskill install --repo anton-abyzov/specweave --plugin sw

# Install domain plugins from the vskill marketplace
npx vskill install --repo anton-abyzov/vskill --plugin mobile
npx vskill install --repo anton-abyzov/vskill --plugin marketing
npx vskill install --repo anton-abyzov/vskill --plugin google-workspace
npx vskill install --repo anton-abyzov/vskill --plugin productivity
npx vskill install --repo anton-abyzov/vskill --plugin skills

Browse 100,000+ community skills at verified-skill.com:

npx vskill find "react"
npx vskill install <skill-name>

Next Steps​