Comprehensive framework for creating, analyzing, and optimizing playbooks that orchestrate skill sequences. Use when creating new playbooks, improving existing playbook workflows, or applying evidence-based orchestration patterns. Applies 7-dimensional playbook analysis, dependency optimization, and parallelization strategies to transform playbooks into highly effective orchestration sequences.
Inherits all available tools
Additional assets for this skill
This skill inherits all available tools. When active, it can use any tool Claude has access to.
README.mdname: playbook-architect description: Comprehensive framework for creating, analyzing, and optimizing playbooks that orchestrate skill sequences. Use when creating new playbooks, improving existing playbook workflows, or applying evidence-based orchestration patterns. Applies 7-dimensional playbook analysis, dependency optimization, and parallelization strategies to transform playbooks into highly effective orchestration sequences. version: 1.0.0 category: foundry tags:
NEVER:
self_consistency: "After playbook creation, trace execution path with multiple test intents to verify consistent routing and skill activation" program_of_thought: "Decompose playbook design into: 1) Intent analysis, 2) Skill mapping, 3) Dependency detection, 4) Parallelization optimization, 5) MCP requirements, 6) Success criteria definition" plan_and_solve: "Plan: Analyze workflow requirements -> Execute: Design skill sequence with dependencies -> Verify: Test with diverse trigger conditions + edge cases"
<!-- END SKILL SOP IMPROVEMENT -->A comprehensive framework for creating, analyzing, and optimizing playbooks that orchestrate skill sequences. This skill applies the same evidence-based methodology used in prompt-architect and agent-creator to the playbook layer of the system.
Playbooks are orchestrated sequences of skills that execute in response to specific user intents. Unlike individual skills (which do ONE thing well), playbooks coordinate MULTIPLE skills to accomplish complex workflows.
The Playbook Architect provides systematic methods to:
USER INTENT
|
v
PLAYBOOK (orchestration layer)
|
+-- Skill 1 (intent-analyzer) -- ALWAYS FIRST
| |
| +-- Agent(s) spawned by skill
|
+-- Skill 2 (prompt-architect) -- ALWAYS SECOND
| |
| +-- Agent(s) spawned by skill
|
+-- Skill 3-N (domain-specific)
| |
| +-- Agent(s) spawned by skill
|
+-- Skill N+1 (validation) -- ALWAYS LAST
|
+-- Agent(s) spawned by skill
When analyzing or creating playbooks, evaluate across these dimensions:
Strong playbooks have specific trigger conditions that clearly distinguish them from alternatives.
Effective sequences include exactly the skills needed, no more, no less.
Optimized playbooks maximize parallelization while respecting true dependencies.
Robust playbooks declare MCP requirements and handle unavailability gracefully.
Production-ready playbooks have explicit, measurable success criteria.
Resilient playbooks anticipate failures and have recovery strategies.
Well-scoped playbooks are neither too simple nor too complex for their purpose.
Objective: Understand the workflow being orchestrated.
Activities:
Intent Definition
Skill Inventory
Dependency Mapping
Validation Gate:
Outputs:
Objective: Design the playbook structure.
Activities:
Sequence Design
playbook: [playbook-name]
version: 1.0.0
triggers:
keywords: [list of trigger keywords]
patterns: [regex patterns]
negative: [keywords that exclude this playbook]
phases:
- name: "Phase 1: [Name]"
type: sequential
skills:
- skill: intent-analyzer
purpose: "Analyze user intent"
output: intent_analysis
- skill: prompt-architect
purpose: "Optimize request"
input: intent_analysis
output: optimized_request
- name: "Phase 2: [Name]"
type: parallel
prerequisites: [Phase 1]
skills:
- skill: [skill-1]
purpose: "[Purpose]"
agent: [agent-name]
- skill: [skill-2]
purpose: "[Purpose]"
agent: [agent-name]
- name: "Phase 3: [Name]"
type: sequential
prerequisites: [Phase 2]
skills:
- skill: [validation-skill]
purpose: "Validate output"
Parallelization Strategy
MCP Configuration
mcp_requirements:
required: [memory-mcp, sequential-thinking]
optional: [playwright, ruv-swarm]
fallbacks:
playwright: manual-testing-skill
Validation Gate:
Outputs:
Objective: Define what success looks like and what to avoid.
Activities:
Success Metrics
success_criteria:
primary_metric: "[What defines success]"
secondary_metrics:
- execution_time: "<[time]"
- quality_score: ">[threshold]"
- test_coverage: ">[percentage]"
acceptance_gates:
- "[Gate 1]: [Condition]"
- "[Gate 2]: [Condition]"
definition_of_done:
- "[Criterion 1]"
- "[Criterion 2]"
- "[Criterion 3]"
Guardrails Definition
guardrails:
NEVER:
- "[Anti-pattern 1 with rationale]"
- "[Anti-pattern 2 with rationale]"
ALWAYS:
- "[Required behavior 1]"
- "[Required behavior 2]"
failure_modes:
- trigger: "[Condition]"
action: "[Recovery step]"
escalate_to: "[human/agent]"
Edge Case Handling
edge_cases:
- case: "[Edge case 1]"
detection: "[How to detect]"
handling: "[What to do]"
- case: "[Edge case 2]"
detection: "[How to detect]"
handling: "[What to do]"
Validation Gate:
Outputs:
Objective: Validate playbook works in practice.
Activities:
Trigger Testing
Execution Testing
Failure Injection
Documentation
Validation Gate:
Outputs:
Purpose: Ship features and functionality Common Skills: sparc-methodology, feature-dev-complete, smart-bug-fix Pattern: Research -> Plan -> Implement -> Test -> Deploy
Purpose: Investigate and analyze Common Skills: deep-research-orchestrator, literature-synthesis, gemini-search Pattern: Question -> Search -> Synthesize -> Validate -> Report
Purpose: Infrastructure and deployment Common Skills: cicd-intelligent-recovery, deployment-readiness, kubernetes-specialist Pattern: Prepare -> Deploy -> Validate -> Monitor -> Recover
Purpose: Audit and improve code quality Common Skills: clarity-linter, code-review-assistant, theater-detection-audit Pattern: Collect -> Analyze -> Report -> Fix -> Verify
Purpose: Security analysis and hardening Common Skills: reverse-engineering-quick-triage, compliance, security-audit Pattern: Assess -> Scan -> Analyze -> Remediate -> Validate
When improving existing playbooks, apply this template:
## Playbook: [Name]
### Trigger Conditions
**When to Use**:
- [Specific trigger 1]
- [Specific trigger 2]
**When NOT to Use**:
- [Negative trigger 1]
- [Negative trigger 2]
### Skill Sequence
| Phase | Type | Skills | Purpose | Prerequisites |
|-------|------|--------|---------|---------------|
| 1 | Sequential | intent-analyzer, prompt-architect | Intent + Optimization | None |
| 2 | Parallel | [skill-a], [skill-b], [skill-c] | [Purposes] | Phase 1 |
| 3 | Sequential | [validation-skill] | Validation | Phase 2 |
### Success Criteria
- **Primary**: [Measurable outcome]
- **Acceptance Gates**: [List of gates]
- **Definition of Done**: [Clear completion criteria]
### Edge Cases
| Case | Detection | Handling |
|------|-----------|----------|
| [Case 1] | [Detection method] | [Recovery action] |
| [Case 2] | [Detection method] | [Recovery action] |
### Guardrails
**NEVER**:
- [Anti-pattern 1]
- [Anti-pattern 2]
**ALWAYS**:
- [Required behavior 1]
- [Required behavior 2]
### MCP Requirements
- **Required**: [List]
- **Optional**: [List]
- **Fallbacks**: [Mapping]
### Complexity and Timing
- **Complexity**: [Low/Medium/High/Very High]
- **Estimated Time**: [Duration]
- **Agent Count**: [Number]
To ensure complete playbook coverage:
List All User Intent Categories
Map Existing Playbooks to Categories
Design Missing Playbooks
Validate MECE Property
Playbook Architect provides the systematic methodology for creating and optimizing playbooks at the same quality level as skills and agents. By applying 7-dimensional analysis, 4-phase creation methodology, and evidence-based patterns, this skill ensures playbooks are:
Use this skill whenever creating new playbooks or improving existing ones to maintain consistency and quality across the entire playbook ecosystem.