Wolf Agents Master Skill
This is the master skill that provides access to Wolf Agents' institutional knowledge accumulated over 50+ phases of development. Enhanced with Superpowers Skill-Chaining patterns (Phase 1 & 2) for dramatically improved agent compliance.
What's New in v1.1.0
Phase 1 & 2 Enhancements (November 2025):
- ✅ Explicit Skill Chaining: "REQUIRED NEXT SKILL" callouts forcing sequential workflows
- ✅ Rationalization Blocking: "Red Flags - STOP" sections catching common shortcuts
- ✅ Mandatory Verification: Checklists with pass/fail criteria that cannot be skipped
- ✅ Good/Bad Examples: Concrete compliance patterns (8 example pairs added)
- ✅ Subagent Templates: 4 role templates for easy delegation
- ✅ Extended Coverage: All core skills + wolf-verification integrated
Expected Impact: Agent compliance rates improved from 30-50% to 90-95% across all governance requirements.
Available Wolf Skills
Core Framework Skills (v1.1.0 - v1.2.0)
🚀 wolf-session-init (v1.0.0) START HERE
Master initialization skill with mandatory 4-step protocol
Use when:
- Starting ANY new work or session
- Session recovery after context loss
- Beginning implementation work
Provides:
- BLOCKING gates for principles → archetype → governance → role
- Session initialization checklist
- Context recovery protocol
MCP Tool: Start with this skill, then follow its chain
🎯 wolf-principles (v1.1.0)
Wolf's 10 core principles guiding system design and agent behavior
Use when:
- Making architectural decisions
- Justifying design choices
- Resolving conflicts between priorities
- Understanding Wolf's philosophy
Enhancements:
- Red Flags - STOP section (6 rationalizations)
- Chains to wolf-archetypes
- Verification checklist (5 items)
MCP Tool: mcp__wolf-knowledge__query_principles({ principle_id: 1-10 })
🔄 wolf-archetypes (v1.2.0)
Behavioral archetype selection with overlay lenses
Use when:
- Starting new work items
- Determining priorities and evidence requirements
- Applying specialized quality gates (performance, security, accessibility, observability)
Enhancements:
- Red Flags - STOP section (7 rationalizations)
- 5 Good/Bad example pairs showing proper archetype selection
- Chains to wolf-governance
- Verification checklist (6 items)
MCP Tool: mcp__wolf-knowledge__find_archetype({ labels: [...], description: "..." })
🛡️ wolf-governance (v1.2.0)
Compliance rules, quality gates, and process standards
Use when:
- Checking Definition of Done requirements
- Understanding quality gates (Fast-Lane, Full-Suite)
- Validating PR readiness
- Understanding approval requirements
Enhancements:
- Red Flags - STOP section (8 rationalizations)
- 3 Good/Bad example pairs (feature PR, security change, refactoring)
- Chains to wolf-roles and wolf-verification
- Verification checklist (8 items)
MCP Tool: mcp__wolf-knowledge__search_governance({ query: "quality gates" })
📋 wolf-roles (v1.2.0)
Guidance for 50+ specialized agent roles with responsibilities
Use when:
- Understanding role responsibilities and boundaries
- Determining collaboration patterns
- Identifying escalation paths
- Using subagent templates for delegation
Enhancements:
- Red Flags - STOP section (7 rationalizations)
- 4 subagent templates (coder, pm, security, code-reviewer)
- Subagent delegation patterns
- Verification checklist (6 items)
- Marks completion of primary skill chain ✅
MCP Tool: mcp__wolf-knowledge__get_role_guidance({ role_name: "agent-role" })
✅ wolf-verification (v1.1.0)
Three-layer verification (CoVe, HSP, RAG) for continuous validation
Use when:
- During implementation at checkpoints
- Before claiming work complete
- Validating evidence requirements
- Checking confidence scores
Enhancements:
- Red Flags - STOP section (7 rationalizations)
- 2 Good/Bad examples
- Integration with governance gates
- Verification checklist (6 items)
MCP Tool: mcp__wolf-core-ip__check_confidence({ model_confidence, evidence_count, ... })
Automation & Scripts (v1.1.0)
⚙️ wolf-scripts-core (v1.1.0)
Core automation for archetype selection, evidence validation, quality scoring
Use when:
- Automating archetype selection
- Validating evidence requirements
- Scoring issue quality with curator rubric
- Validating bash scripts
Enhancements:
- Red Flags - STOP section (6 rationalizations)
- 2 Good/Bad examples (archetype selection, evidence validation)
- Integration with wolf-archetypes and wolf-governance
- Verification checklist (5 items)
Scripts: select-archetype.mjs, evidence-validator.mjs, curator-rubric.mjs, bash-validator.mjs
🤖 wolf-scripts-agents (v1.1.0)
Agent coordination, orchestration, and multi-agent workflow management
Use when:
- Coordinating multi-agent workflows
- Enforcing agent file scope boundaries
- Using mailbox system for async communication
- Orchestrating complex pipelines
Enhancements:
- Red Flags - STOP section (6 rationalizations)
- 2 Good/Bad examples (workflow orchestration, scope validation)
- Integration with wolf-roles
- Verification checklist (6 items)
Scripts: orchestrate-workflow.mjs, validate-agent-changes.mjs, agent-executor.mjs
Supporting Skills (v1.0.0)
📚 wolf-instructions
Four-level instruction cascading (Global → Domain → Project → Role)
Use when:
- Resolving instruction conflicts
- Understanding priority hierarchy
- Loading contextual guidance
📝 wolf-adr
Architecture Decision Records system
Use when:
- Documenting architectural decisions
- Understanding past decisions
- Creating ADRs for major changes
Complete Skill Chain Diagram
SESSION START
|
v
[wolf-session-init] - MANDATORY ENTRY POINT
| Step 1: Query principles (BLOCKING)
| Step 2: Find archetype (BLOCKING)
| Step 3: Load governance (BLOCKING)
| Step 4: Load role (BLOCKING)
v
PRIMARY SKILL CHAIN (Sequential - DO NOT skip)
|
v
[wolf-principles] (v1.1.0)
| → Strategic guidance and decision framework
| → REQUIRED NEXT: wolf-archetypes
v
[wolf-archetypes] (v1.2.0)
| → Behavioral profile and evidence requirements
| → Apply lenses if needed (performance, security, accessibility, observability)
| → REQUIRED NEXT: wolf-governance
v
[wolf-governance] (v1.2.0)
| → Definition of Done, quality gates, compliance
| → REQUIRED NEXT: wolf-roles
| → REQUIRED ALWAYS: wolf-verification
v
[wolf-roles] (v1.2.0)
| → Role responsibilities, collaboration patterns
| → Use subagent templates for delegation
| → PRIMARY CHAIN COMPLETE ✅
v
IMPLEMENTATION BEGINS
|
v
[wolf-verification] (v1.1.0) - Called DURING work at checkpoints
| → Three-layer validation (CoVe, HSP, RAG)
| → Evidence collection
| → Confidence scoring
v
COMPLETION ✅When to Use Each Skill - Decision Tree
START: Are you beginning new work or recovering context?
│
├─ YES → Use wolf-session-init (MANDATORY)
│ └─ Follow its 4-step blocking protocol
│
└─ NO: What do you need?
│
├─ "Strategic guidance / decision framework"
│ └─ Use wolf-principles
│ └─ MCP: mcp__wolf-knowledge__query_principles
│
├─ "What archetype for this work type?"
│ └─ Use wolf-archetypes
│ └─ MCP: mcp__wolf-knowledge__find_archetype
│ └─ OR Script: select-archetype.mjs
│
├─ "What are quality gates / Definition of Done?"
│ └─ Use wolf-governance
│ └─ MCP: mcp__wolf-knowledge__search_governance
│
├─ "What are my role responsibilities?"
│ └─ Use wolf-roles
│ └─ MCP: mcp__wolf-knowledge__get_role_guidance
│ └─ Use templates for subagent delegation
│
├─ "How do I validate evidence / check con