specweave 0.1.9 → 0.3.1

package/src/skills/context-optimizer/SKILL.md

@@ -354,37 +354,7 @@ You can now work on payment features with optimized context.
 
 ## Configuration
 
-```yaml
-# .specweave/config.yaml
-context_optimizer:
-  enabled: true
-
-  # Auto-run after context-loader
-  auto_optimize: true
-
-  # Minimum context size to trigger optimization
-  min_context_tokens: 20000
-
-  # Confidence threshold (skip if uncertain)
-  min_confidence: 0.75
 
-  # Show what was removed
-  show_removals: true
-
-  # Keep buffer (don't remove if close to needed)
-  buffer_strategy: "conservative"  # or "aggressive"
-
-  # Domain detection
-  custom_domains:
-    - "real-time-chat"
-    - "analytics"
-    - "reporting"
-
-  # Always keep (never remove)
-  always_keep:
-    - "architecture/README.md"
-    - "CLAUDE.md"
-```
 
 ## Integration with Context Loader
 

package/src/skills/github-sync/SKILL.md

@@ -36,25 +36,7 @@ description: Bi-directional synchronization between GitHub and SpecWeave. Maps G
 
 ## Configuration
 
-**.specweave/config.yaml**:
-```yaml
-sync:
-  github:
-    enabled: true
-    repo: company/repo
-    auth:
-      type: token
-      token_env: GITHUB_TOKEN
-    sync_interval: 15m
-    webhooks:
-      enabled: true
-      secret_env: GITHUB_WEBHOOK_SECRET
-    mappings:
-      milestone: release_plan
-      issue_feature: rfc  # Issues labeled "feature"
-      issue_bug: incident  # Issues labeled "bug"
-      issue_task: task     # Issues labeled "task"
-```
+GitHub sync uses auto-detection - no configuration file needed. Repository and credentials are detected from environment variables or prompted when needed.
 
 ---
 

package/src/skills/increment-planner/SKILL.md

@@ -68,27 +68,73 @@ STEP 1: Scan existing docs
 ├─ Read .specweave/docs/internal/architecture/adr/ (existing decisions)
 └─ Pass existing context to agents
     ↓
-STEP 2: Invoke PM Agent
-├─ PM creates .specweave/docs/internal/strategy/crypto-trading/
-│   ├── overview.md (product vision, problem, users)
-│   ├── requirements.md (FR-001, NFR-001, technology-agnostic)
-│   ├── user-stories.md (US1, US2, US3...)
-│   └── success-criteria.md (KPIs, metrics)
-└─ PM creates .specweave/increments/0001-*/spec.md (summary, references strategy docs)
+STEP 2: Invoke PM Agent (🚨 MANDATORY - USE TASK TOOL)
+
+YOU MUST USE THE TASK TOOL - DO NOT SKIP:
+
+Task(
+  subagent_type: "pm",
+  description: "PM product strategy",
+  prompt: "Create product strategy for: [user feature description]
+
+  Context from existing docs: [summary of strategy docs from Step 1]
+
+  You MUST create BOTH living docs AND increment files:
+
+  1. Living docs (source of truth):
+     - Create .specweave/docs/internal/strategy/{module-name}/
+       * overview.md (product vision, problem, users)
+       * requirements.md (FR-001, NFR-001, technology-agnostic)
+       * user-stories.md (US1, US2, US3...)
+       * success-criteria.md (KPIs, metrics)
+
+  2. Increment file:
+     - Create .specweave/increments/{number}-{name}/spec.md
+     - Keep spec.md < 250 lines (summary only)
+     - MUST reference living docs
+     - Include links to ../../docs/internal/strategy/{module}/
+
+  Tech stack: [detected tech stack]"
+)
+
+Wait for PM agent to complete!
     ↓
-STEP 3: Invoke Architect Agent
-├─ Architect READS strategy docs from PM
-├─ Architect creates .specweave/docs/internal/architecture/
-│   ├── system-design.md (overall architecture)
-│   ├── adr/0001-websocket-vs-polling.md (decision record)
-│   ├── adr/0002-database-choice.md (PostgreSQL vs MongoDB)
-│   ├── adr/0003-deployment-platform.md (Railway vs Hetzner)
-│   └── diagrams/crypto-trading/ (Mermaid C4 diagrams)
-└─ Architect creates .specweave/increments/0001-*/plan.md (summary, references architecture docs)
+STEP 3: Invoke Architect Agent (🚨 MANDATORY - USE TASK TOOL)
+
+YOU MUST USE THE TASK TOOL - DO NOT SKIP:
+
+Task(
+  subagent_type: "architect",
+  description: "Architect technical design",
+  prompt: "Create technical architecture for: [user feature description]
+
+  FIRST, read PM's strategy docs from .specweave/docs/internal/strategy/{module}/
+
+  You MUST create BOTH living docs AND increment files:
+
+  1. Living docs (source of truth):
+     - Update .specweave/docs/internal/architecture/system-design.md
+     - Create .specweave/docs/internal/architecture/adr/ (at least 3 ADRs):
+       * ####-websocket-vs-polling.md
+       * ####-database-choice.md
+       * ####-deployment-platform.md
+     - Create diagrams/{module}/ (Mermaid C4 diagrams)
+     - Create data-models/{module}-schema.sql
+
+  2. Increment file:
+     - Create .specweave/increments/{number}-{name}/plan.md
+     - Keep plan.md < 500 lines (summary only)
+     - MUST reference living docs
+     - Include links to ../../docs/internal/architecture/adr/
+
+  Tech stack: [detected tech stack]"
+)
+
+Wait for Architect agent to complete!
     ↓
 STEP 4: Validate Living Docs
 ├─ Check .specweave/docs/internal/strategy/{module}/ is NOT empty
-├─ Check .specweave/docs/internal/architecture/adr/ has ADRs
+├─ Check .specweave/docs/internal/architecture/adr/ has ≥3 ADRs
 ├─ Check increment spec.md REFERENCES (not duplicates) strategy docs
 └─ Check increment plan.md REFERENCES (not duplicates) architecture docs
     ↓
@@ -824,7 +870,7 @@ const featurePlanner = require('./create-feature.js');
 const feature = await featurePlanner.createFeature({
   description: "Add Stripe payment integration",
   priority: "P1",
-  configPath: ".specweave/config.yaml"
+  # Configuration auto-detected
 });
 
 console.log(`Created: .specweave/increments/${feature.number}-${feature.shortName}/`);

package/src/skills/increment-quality-judge/SKILL.md

@@ -238,42 +238,7 @@ Choice: _
 
 ## Configuration
 
-```yaml
-# .specweave/config.yaml
-validation:
-  quality_judge:
-    # Enable/disable feature
-    enabled: true
-
-    # When to run
-    auto_prompt: true        # Prompt after rule-based validation
-    always_run: false        # Run without prompting
-
-    # Scoring thresholds
-    thresholds:
-      excellent: 90          # ✓✓ (95-100)
-      good: 80               # ✓  (80-94)
-      acceptable: 70         # ~  (70-79)
-      needs_work: 0          # ⚠️ (0-69)
-
-    # Which dimensions to evaluate
-    dimensions:
-      clarity: true
-      testability: true
-      completeness: true
-      feasibility: true
-      maintainability: true
-      edge_cases: true
-
-    # Token budget
-    max_tokens: 2000         # Limit for quality check
-
-    # Auto-export suggestions
-    export_to_tasks: true    # Add suggestions to tasks.md
-
-    # Confidence threshold
-    min_confidence: 0.80     # Show warning if lower
-```
+
 
 ## Usage Examples
 

package/src/skills/jira-sync/README.md

@@ -91,15 +91,7 @@ As a user, I want to subscribe to a monthly plan...
 ## Authentication
 
 **JIRA Cloud**:
-```yaml
-# .specweave/config.yaml
-jira_sync:
-  type: "cloud"
-  url: "https://your-org.atlassian.net"
-  email: "user@example.com"
-  api_token: "${JIRA_API_TOKEN}"  # From environment variable
-  project_key: "PROJ"
-```
+
 
 **JIRA Server/Data Center**:
 ```yaml
@@ -113,36 +105,7 @@ jira_sync:
 
 ## Configuration
 
-```yaml
-# .specweave/config.yaml
-jira_sync:
-  enabled: false
-
-  # JIRA instance
-  url: "https://your-org.atlassian.net"
-  project_key: "PROJ"
-
-  # Authentication (use environment variables!)
-  email: "${JIRA_EMAIL}"
-  api_token: "${JIRA_API_TOKEN}"
-
-  # Sync settings
-  sync_direction: "bidirectional"  # export | import | bidirectional
-  sync_interval: 300  # seconds (5 minutes)
 
-  # What to sync
-  sync_status: true
-  sync_comments: true
-  sync_assignees: true
-  sync_attachments: false
-
-  # Issue mapping
-  epic_label: "epic"  # Custom field or label for epics
-
-  # Webhooks (for real-time sync)
-  webhook_enabled: true
-  webhook_secret: "${JIRA_WEBHOOK_SECRET}"
-```
 
 ## Workflow
 

package/src/skills/role-orchestrator/README.md

@@ -160,28 +160,7 @@ const roleRules = {
 
 ## Configuration
 
-```yaml
-# .specweave/config.yaml
-role_orchestrator:
-  enabled: true
-
-  # Model selection per role
-  models:
-    pm: "claude-opus-4"
-    architect: "claude-opus-4"
-    security: "claude-opus-4"
-    qa_lead: "claude-sonnet-4"
-    devops: "claude-sonnet-4"
-    tech_lead: "claude-sonnet-4"
-    backend: "claude-haiku-4"
-    frontend: "claude-haiku-4"
-
-  # Auto-detect roles or ask user
-  auto_detect: true
-
-  # Ask user confirmation before running expensive roles (Opus)
-  confirm_opus_roles: true
-```
+
 
 ## Future Enhancements
 

package/src/skills/role-orchestrator/SKILL.md

@@ -889,37 +889,7 @@ Proceeding to architecture phase...
 
 ### Configuration for Feedback Loops
 
-```yaml
-# .specweave/config.yaml
-role_orchestrator:
-  enabled: true
-
-  # Feedback loop configuration
-  feedback_loops:
-    enabled: true              # Enable auto-refinement
-    max_retries: 3             # Max refinement attempts
-    stop_on_improvement: true  # Stop if score improves
-    require_user_approval: false  # Auto-refine without asking
-
-    # Quality thresholds per agent
-    thresholds:
-      pm_agent: 0.80           # Requirements quality
-      architect_agent: 0.80    # Design quality
-      qa_lead_agent: 0.75      # Test coverage + quality
-
-    # Which agents use feedback loops
-    agents:
-      - pm_agent
-      - architect_agent
-      - qa_lead_agent
-      # Not applicable for implementation agents
-
-    # Validation strategy
-    validation:
-      use_llm_judge: true      # Use increment-quality-judge
-      combine_with_rules: true # Combine with rule-based checks
-      judge_weight: 0.5        # 50% LLM judge, 50% rules
-```
+
 
 ### Gate Failure Handling
 
@@ -1044,35 +1014,7 @@ async function invokeSkill(skillName: string, task: string, context: any) {
 
 ## Configuration
 
-```yaml
-# .specweave/config.yaml
-role_orchestrator:
-  enabled: true
-
-  # Orchestration strategy
-  default_pattern: "sequential_with_gates"  # or "parallel" or "adaptive"
 
-  # Quality gates
-  require_user_approval:
-    - architecture_decisions
-    - deployment_to_production
-
-  require_automated_approval:
-    - test_coverage: ">80%"
-    - security_scan: "no_critical"
-
-  # Agent preferences
-  preferred_backend: "nodejs-backend"  # or python-backend, dotnet-backend
-  preferred_frontend: "frontend-agent"
-
-  # Timeouts
-  phase_timeout_minutes: 30
-  total_timeout_hours: 8
-
-  # Monitoring
-  progress_updates_interval: "5min"
-  store_metrics: true
-```
 
 ## Testing
 

package/src/skills/skill-router/SKILL.md

@@ -285,25 +285,7 @@ When multiple skills could handle a request, use priority:
 
 ## Configuration
 
-```yaml
-# .specweave/config.yaml
-skill_router:
-  enabled: true
 
-  # Accuracy target (constitution requirement)
-  accuracy_target: 0.90
-
-  # Clarification threshold
-  clarification_threshold: 0.90  # Ask when confidence < 0.90
-
-  # Learning
-  collect_feedback: true
-  feedback_file: ".specweave/cache/routing-feedback.json"
-
-  # Routing preferences
-  prefer_multi_skill: false  # For complex requests, prefer orchestrator
-  auto_route_high_confidence: true  # Route immediately if confidence > 0.90
-```
 
 ## Examples
 

package/src/skills/spec-kit-expert/SKILL.md

@@ -356,7 +356,6 @@ When you request gap analysis, I will:
 ```bash
 # Project structure
 - CLAUDE.md (complete guide)
-- .specweave/config.yaml (configuration)
 - .specweave/increments/ (implemented features)
 - src/skills/ (available skills)
 - src/agents/ (available agents)
@@ -436,7 +435,7 @@ Based on analysis, provide:
    - Read src/skills/ for available skills
    - Read src/agents/ for available agents
    - Review CLAUDE.md for documented features
-   - Check .specweave/config.yaml for configuration
+   - Settings auto-detected
    - Analyze .claude/hooks/ for automation
    ```
 
@@ -885,7 +884,6 @@ When performing gap analysis, I can read and analyze:
 ```bash
 # Core configuration
 ✅ CLAUDE.md                          # Complete guide
-✅ .specweave/config.yaml            # Configuration
 ✅ README.md                         # Project overview
 
 # Features (increments)