Skip to main content

🧩 Plugin Ecosystem

SpecWeave's power comes from its modular plugin architecture. With 21 specialized plugins, you get enterprise-grade capabilities that activate automatically based on your project needs.

Key Insight

Progressive Disclosure: Only relevant plugins activate based on context. Working on React? Frontend plugin loads. Mention Kubernetes? Infrastructure plugin activates. This keeps your AI context lean and efficient.

🎯 Quick Start​

All plugins are automatically installed during specweave init:

npx specweave init .
# ✅ Installs all 21 plugins (GitHub, JIRA, K8s, React, ML, etc.)
# ✅ Registers marketplace globally
# ✅ Ready to use immediately!

📊 Plugin Categories​

🔧 Core Framework (Always Active)​

sw - Essential Framework​

The foundation of SpecWeave, always loaded in every project.

Key Capabilities:

  • 🎯 9 Skills: increment-planner, spec-generator, tdd-workflow, brownfield-analyzer, context-optimizer
  • 🤖 22 Agents: PM, Architect, Tech Lead, QA Lead, Security Engineer, Code Reviewer
  • ⚡ 22 Commands: Full increment lifecycle management

Example: Planning a New Feature

# The PM agent (from core plugin) helps you create a spec
/sw:increment "User authentication with OAuth"

# Output: Creates spec.md, plan.md, tasks.md with:
# - User stories with acceptance criteria
# - Technical design with architecture
# - Task breakdown with embedded test plans
# - [BDD](/docs/glossary/terms/bdd) format (Given/When/Then)

Skills Highlight:

  • increment-planner: Automatically activates when you say "plan feature", "new increment", "build X"
  • tdd-workflow: Activates on "TDD", "test-driven", "red-green-refactor"
  • brownfield-analyzer: Detects existing projects and maps documentation to SpecWeave structure

🔗 External Integrations​

sw-github - GitHub Issues Integration​

Bidirectional sync between SpecWeave increments and GitHub Issues.

Key Features:

  • ✅ Auto-creates issues on increment planning
  • ✅ Task-level progress tracking (checkboxes)
  • ✅ Universal Hierarchy: Epic → Milestone, Increment → Issue
  • ✅ Smart conflict resolution

Example: Auto-Sync Workflow

# 1. Create increment (issue auto-created!)
/sw:increment "Add dark mode"
# 🔗 GitHub Issue #42 created automatically

# 2. Start work
/sw:do

# 3. Complete a task
# ✅ GitHub issue checkbox updates automatically!

# 4. Manual sync (if needed)
/sw-github:sync 0023 --time-range 1M

Commands:

  • /sw-github:sync - Bidirectional sync with time range filtering
  • /sw-github:create-issue - Manual issue creation
  • /sw-github:sync-epic - Sync Epic folder to GitHub Milestone
  • /sw-github:status - Check sync status

Configuration:

{
  "sync": {
    "settings": {
      "autoCreateIssue": true,
      "syncDirection": "bidirectional"
    },
    "profiles": {
      "frontend": {
        "provider": "github",
        "config": {
          "owner": "myorg",
          "repo": "frontend"
        }
      }
    }
  }
}

sw-jira - JIRA Integration​

Enterprise JIRA integration with Epic/Story sync.

Key Features:

  • ✅ Bidirectional Epic ↔ JIRA Epic sync
  • ✅ User Stories ↔ JIRA Stories
  • ✅ Status mapping (Planning → To Do, Active → In Progress)
  • ✅ Multi-project support

Example: JIRA Sync

# Sync increment to JIRA
/sw-jira:sync 0042

# Sync Epic folder (20+ user stories)
/sw-jira:sync-epic FS-031-authentication

# Check status
/sw-jira:status 0042

Status Mapping:

SpecWeave          → JIRA
─────────────────────────────
planning           → To Do
active             → In Progress
paused             → On Hold
completed          → Done
abandoned          → Cancelled

sw-ado - Azure DevOps Integration​

Enterprise Azure DevOps sync with Work Items.

Key Features:

  • ✅ Epic → Azure DevOps Epic
  • ✅ Increment → User Story
  • ✅ Tasks → Tasks with Area Paths
  • ✅ Multi-project organization strategies

Example:

# Sync to Azure DevOps
/sw-ado:sync 0031

# Create work item manually
/sw-ado:create-workitem 0031

# Check sync status
/sw-ado:status 0031

sw-figma - Figma Design Integration​

Connect designs to implementation.

Skills:

  • figma-integration: Extract design tokens, components, screens
  • Converts Figma → React components
  • Auto-generates prop types from Figma properties

💻 Tech Stack Plugins​

sw-frontend - Frontend Development​

React, Vue, Angular, Next.js expertise.

Agents:

  • frontend-architect: Component architecture, state management patterns
  • react-expert: Hooks, performance, modern patterns
  • nextjs-specialist: SSR, ISR, routing, API routes

Example: React Component Planning

# Frontend architect activates automatically
"I need to build a reusable data table component with sorting and filtering"

# Generates:
# - Component architecture (container/presentational)
# - State management approach (useState, useReducer, Zustand)
# - Performance optimizations (React.memo, useMemo)
# - Accessibility (ARIA labels, keyboard navigation)
# - Test strategy (Jest, React Testing Library)

Tech Stack Coverage:

  • React, Vue 3, Angular, Svelte
  • Next.js, Nuxt, SvelteKit
  • State Management: Redux, Zustand, Jotai, Pinia
  • Styling: Tailwind CSS, Styled Components, Emotion

sw-backend - Backend Development​

Node.js, Python, .NET expertise.

Agents:

  • backend-architect: API design, microservices, authentication
  • nodejs-expert: Express, Fastify, NestJS patterns
  • python-expert: Django, FastAPI, async patterns

Example: API Design

"Design RESTful API for user management with authentication"

# Backend architect generates:
# - OpenAPI/Swagger spec
# - Authentication strategy (JWT, OAuth2)
# - Database schema (PostgreSQL)
# - Rate limiting and caching
# - Error handling patterns
# - API versioning strategy

sw-mobile - Mobile Development​

React Native and Expo expertise.

Agents:

  • mobile-architect: Navigation, state, offline-first patterns
  • react-native-expert: Native modules, performance optimization

Example:

"Build offline-first mobile app with background sync"

# Mobile architect provides:
# - React Native architecture
# - Offline data strategy (SQLite, Realm, WatermelonDB)
# - Background sync with NetInfo
# - Push notifications setup
# - Platform-specific considerations (iOS/Android)

☁️ Infrastructure & DevOps​

sw-infra - Infrastructure Engineering​

Cloud infrastructure with Hetzner, DigitalOcean, monitoring.

Agents:

  • devops: CI/CD pipelines, cloud deployments
  • sre: Incident response, troubleshooting, root cause analysis
  • observability-engineer: Monitoring, logging, distributed tracing
  • network-engineer: VPC, firewalls, load balancers

Example: Production Incident

"API response time increased to 5 seconds, CPU at 90%"

# SRE agent activates with:
# - Incident severity classification (SEV2)
# - Immediate mitigation steps
# - Root cause analysis checklist
# - Performance profiling strategy
# - Post-mortem template

Playbooks:

  • High CPU Usage
  • Database Deadlock
  • Memory Leak Detection
  • Slow API Response
  • DDoS Attack Mitigation
  • Disk Full Recovery
  • Service Down Recovery
  • Data Corruption
  • Cascade Failure Prevention
  • Rate Limit Exceeded

sw-k8s - Kubernetes Orchestration​

Cloud-native Kubernetes architecture, GitOps, service mesh.

Agent:

  • kubernetes-architect: EKS/AKS/GKE, Helm, ArgoCD/Flux, Istio

Example: K8s Setup

"Deploy microservices to Kubernetes with GitOps"

# K8s architect provides:
# - Helm chart structure
# - ArgoCD/Flux configuration
# - Service mesh setup (Istio)
# - Multi-tenancy strategy
# - Monitoring with Prometheus/Grafana
# - Security policies (Network Policies, PSP)

🤖 Machine Learning & Data Science​

sw-ml - ML Engineering​

Complete ML lifecycle from data to deployment.

Agents:

  • ml-engineer: Model training, deployment, MLOps
  • data-scientist: Statistical analysis, predictive modeling, EDA
  • mlops-engineer: MLflow, Kubeflow, experiment tracking

Commands:

  • /sw-ml:pipeline - Design complete ML pipeline
  • /sw-ml:evaluate - Evaluate model with metrics
  • /sw-ml:deploy - Generate deployment artifacts
  • /sw-ml:explain - Model explainability (SHAP, LIME)

Example: ML Pipeline

# Design complete ML workflow
/sw-ml:pipeline "Customer churn prediction"

# Generates:
# 1. Data preprocessing (feature engineering, scaling)
# 2. Model training (XGBoost, Random Forest, Neural Network)
# 3. Hyperparameter tuning (Optuna)
# 4. Model evaluation (precision, recall, F1, ROC-AUC)
# 5. MLflow experiment tracking
# 6. Model registry and versioning
# 7. Deployment (FastAPI, Docker, Kubernetes)
# 8. Monitoring (data drift, model performance)

Tech Stack:

  • Frameworks: TensorFlow, PyTorch, scikit-learn, XGBoost
  • MLOps: MLflow, Kubeflow, Weights & Biases
  • Deployment: FastAPI, TorchServe, TFServing
  • Monitoring: Evidently AI, WhyLabs

💳 Specialized Domains​

sw-payments - Payment Integration​

Stripe, PayPal, payment processor expertise.

Agent:

  • payment-integration: Checkout flows, webhooks, PCI compliance

Example:

"Implement Stripe subscription billing"

# Payment integration agent provides:
# - Stripe Checkout setup
# - Subscription lifecycle management
# - Webhook handling (payment_intent.succeeded)
# - Invoice generation
# - Failed payment retry logic
# - PCI compliance checklist

📚 Documentation & Diagrams​

sw-docs - Documentation & Preview​

Documentation generation, organization, and live Docusaurus preview with hot reload.

Commands:

  • /sw-docs:view - Launch interactive docs server (internal or public)
  • /sw-docs:build - Build static site for deployment
  • /sw-docs:generate - Generate documentation
  • /sw-docs:organize - Organize large doc folders
  • /sw-docs:health - Documentation health report
  • /sw-docs:validate - Validate documentation (MDX, YAML, links)

Example:

# View internal docs (default) - port 3015
/sw-docs:view

# View public docs - port 3016
/sw-docs:view --public

# Output:
# 🚀 Server running at http://localhost:3015 (or 3016 for public)
# 📁 Auto-generated sidebar from .specweave/docs/
# 🔄 Hot reload enabled
# 📊 Mermaid diagrams rendered

sw-diagrams - Diagram Generation​

Mermaid diagrams following C4 Model conventions.

Agent:

  • diagrams-architect: C4 Context/Container/Component, sequence, ER diagrams

Example:

"Create C4 Container diagram for microservices architecture"

# Generates:
```mermaid
graph TB
    User[User]
    WebApp[Web Application]
    API[API Gateway]
    AuthService[Auth Service]
    UserService[User Service]
    DB[(Database)]

    User -->|HTTPS| WebApp
    WebApp -->|REST| API
    API -->|gRPC| AuthService
    API -->|gRPC| UserService
    UserService -->|SQL| DB

---

### 🚀 Release Management

#### **sw-release** - Release Orchestration
Multi-repo releases, semantic versioning, RC workflows.

**Agents:**
- **release-manager**: Coordinates releases across monorepo/polyrepo

**Commands:**
- `/sw-release:init` - Initialize release strategy
- `/sw-release:align` - Align versions across repos
- `/sw-release:rc` - Manage Release Candidate lifecycle
- `/sw-release:platform` - Multi-repo platform releases

**Example:**
```bash
# Initialize release strategy
/sw-release:init
# Analyzes git history, CI/CD configs, recommends strategy

# Create Release Candidate
/sw-release:rc create v1.5.0-rc.1

# Promote to production
/sw-release:rc promote v1.5.0-rc.3

🎨 UI Testing & Automation​

sw-testing - Testing & Browser Automation​

Playwright integration with MCP protocol.

Features:

  • Browser automation for E2E testing
  • Visual regression testing
  • Accessibility testing
  • Performance monitoring

🔧 Utilities​

sw-cost - AI Cost Optimization​

Tracks AI costs, suggests optimizations.

Features:

  • Cost tracking per increment
  • Model selection recommendations (Opus vs Haiku)
  • Context optimization suggestions

sw-alternatives - Tool Alternatives​

Suggests alternatives when users mention specific tools.

Example:

"I want to use MongoDB"

# Suggests:
# ✅ MongoDB (NoSQL, document store)
# 🔄 Alternatives:
#   - PostgreSQL with JSONB (relational + document)
#   - CouchDB (CouchDB protocol)
#   - RethinkDB (real-time queries)

🎯 How Plugins Work Together​

Example: Full-Stack Feature Implementation​

# 1. PM Agent (Core) creates spec
/sw:increment "Real-time chat feature"

# 2. Architect Agent (Core) designs system
# - WebSocket architecture
# - Database schema (PostgreSQL)
# - Caching strategy (Redis)

# 3. Frontend Agent (Frontend Plugin) designs UI
# - React components
# - State management (Zustand)
# - WebSocket client

# 4. Backend Agent (Backend Plugin) designs API
# - WebSocket server (Socket.io)
# - Authentication (JWT)
# - Message persistence

# 5. K8s Agent (Infrastructure Plugin) plans deployment
# - Helm charts
# - Horizontal Pod Autoscaler
# - Ingress configuration

# 6. GitHub Agent (GitHub Plugin) creates issue
# - Auto-creates GitHub Issue #84
# - Links to increment

# 7. Start implementation
/sw:do

# 8. All agents collaborate during implementation!

📈 Token Efficiency​

Context Optimization in Action:

ScenarioActive PluginsToken Usage
Planning OnlyCore~12K tokens
React FeatureCore + Frontend~18K tokens
Full-Stack APICore + Frontend + Backend~24K tokens
ML PipelineCore + ML + Infrastructure~28K tokens
Enterprise (All)All 21 plugins~30K tokens

Result: 75% smaller than monolithic architecture!

🔍 Finding the Right Plugin​

I want to...Use PluginKey Command
Plan incrementsw/sw:increment
Sync to GitHubsw-github/sw-github:sync
Deploy to K8ssw-k8sAgent auto-activates
Build React appsw-frontendAgent auto-activates
Create ML pipelinesw-ml/sw-ml:pipeline
View docssw-docs/sw-docs:view
Handle incidentssw-infraSRE agent + playbooks
Generate diagramssw-diagramsAgent auto-activates
Integrate Stripesw-paymentsAgent auto-activates
Manage releasessw-release/sw-release:init

🚀 Quick Reference: All 20 Plugins​

PluginPurposeKey Features
swCore framework22 agents, 22 commands, increment lifecycle
sw-githubGitHub syncBidirectional sync, auto-issue creation
sw-jiraJIRA syncEpic/Story sync, status mapping
sw-adoAzure DevOpsWork items, multi-project strategies
sw-frontendFrontend devReact, Vue, Angular, Next.js
sw-backendBackend devNode.js, Python, .NET, APIs
sw-mobileMobile devReact Native, Expo, offline-first
sw-infraDevOpsSRE, monitoring, incident response
sw-k8sK8s orchestrationHelm, GitOps, service mesh
sw-mlMachine LearningMLOps, pipelines, deployment
sw-releaseRelease mgmtSemantic versioning, RC workflow
sw-paymentsPaymentsStripe, PayPal, PCI compliance
sw-testingTesting & automationPlaywright, E2E testing
sw-diagramsDiagram generationMermaid, C4 Model
sw-figmaDesign integrationFigma → Components
sw-costCost trackingAI cost optimization
sw-alternativesTool suggestionsAlternative recommendations
sw-docsDocumentationGuides, API docs, Docusaurus preview
sw-confluentKafka/ConfluentStreaming, Schema Registry
sw-kafkaKafka operationsDevOps, monitoring, architecture

💡 Tips for Maximum Efficiency​

1. Trust Progressive Disclosure​

Don't manually load plugins. Let skills activate based on context:

  • Say "React" → Frontend plugin activates
  • Say "Kubernetes" → Infrastructure plugin activates
  • Say "ML pipeline" → ML plugin activates

2. Use Context Optimizer​

For large projects with 100+ files:

"Optimize context for authentication feature"
# Context optimizer removes 80% irrelevant specs/modules

3. Leverage Command Shortcuts​

# Planning
/sw:increment → /sw:do → /sw:done

# Quality
/sw:validate → /sw:qa → /sw:check-tests

# Sync
/sw-github:sync → /sw:sync-docs

4. Multi-Project Setup​

{
  "sync": {
    "profiles": {
      "frontend": {"provider": "github", "config": {"repo": "frontend"}},
      "backend": {"provider": "github", "config": {"repo": "backend"}},
      "mobile": {"provider": "github", "config": {"repo": "mobile"}}
    }
  }
}

🎓 Learning Resources​

🔧 Extensibility via CLAUDE.md​

SpecWeave is a framework, not a locked product. Beyond plugins, you can customize behavior through your project's CLAUDE.md file:

Custom Sync Rules​

## Sync Customization
When syncing to JIRA, always:
- Add custom field "Team: Backend" for backend increments
- Map "paused" status to "Blocked" instead of "On Hold"
- Include sprint field using our sprint naming convention (SPRINT-YYYY-WW)

Custom Quality Gates​

## Quality Gates
Before closing any increment:
- Run `npm run lint:strict` in addition to tests
- Verify CHANGELOG.md entry exists for the feature
- Check that OpenAPI spec is regenerated if API changed
- Ensure Sentry release is tagged

Custom Workflow Hooks​

## Workflow Automation
After completing a task:
- Post notification to #dev-updates Slack channel
- Update team capacity spreadsheet via webhook
- Trigger downstream pipeline if it's an API change

Agent Behavior Overrides​

## Agent Customization
When the PM agent creates specs:
- Always include compliance section for HIPAA requirements
- Add "Data Classification" field to every user story
- Reference our design system component library

When the Architect agent designs:
- Prefer PostgreSQL over MySQL for new services
- Always include rate limiting in API designs
- Use our standard error response format

What you can customize:

AreaExamples
External SyncAdd fields, transform statuses, integrate with internal tools
Quality GatesCustom validation, linting rules, security scans, compliance checks
Lifecycle HooksTrigger actions on events (created, done, paused, archived)
Agent BehaviorOverride default prompts, add domain-specific requirements
Naming ConventionsEnforce team-specific ID formats, branch names, commit messages
Integration LogicCustom webhook payloads, API transformations, field mappings

🤝 Contributing​

Want to add a new plugin? See CLAUDE.md for plugin development guide.

Plugin Ideas Welcome:

  • Cloud providers (AWS, GCP, Azure)
  • Languages (Go, Rust, Java)
  • Frameworks (Spring Boot, Django, Rails)
  • Databases (MongoDB, Redis, Cassandra)
  • Testing tools (Cypress, Playwright)
Next Steps
  1. ✅ Install: npx specweave init .
  2. 🎯 Plan: /sw:increment "Your feature"
  3. ⚡ Implement: /sw:do
  4. 🔗 Sync: /sw-github:sync
  5. ✅ Complete: /sw:done

All 21 plugins are ready to help you build faster!