deepen-plan

SKILL.md

Deepen Plan Skill

Purpose: Enhance existing plans with parallel research agents for maximum depth and grounding.

Plan to Deepen

Deepen target: ${ARGUMENTS}

Deepening Methodology

Phase 1: Plan Analysis

  1. Read and Understand Plan

    • Read the entire plan document (expected location: docs/plans/)
    • If the plan frontmatter contains source-spec:, read the referenced spec file for additional context
    • Identify key sections and requirements
    • Note areas that need additional depth or clarification
  2. Scope Definition

    • Determine which sections benefit most from deepening
    • Identify knowledge gaps or uncertainties
    • Prioritize based on impact and complexity

Phase 2: Parallel Research Agents

Launch specialized research agents in parallel based on plan needs:

Architecture & Design Agents:

  1. System Architecture Researcher - Overall system design and integration
  2. Domain Model Researcher - Business logic and data relationships
  3. API Design Researcher - Endpoint specifications and contracts

Implementation & Quality Agents: 4. Best Practices Researcher - Industry standards and conventions 5. Performance Researcher - Optimization strategies and patterns 6. Security Researcher - Security considerations and mitigations 7. Testing Strategy Researcher - Comprehensive testing approaches

Language & Framework Agents: 8. Language-Specific Researcher - Language idioms and patterns 9. Framework Researcher - Framework capabilities and constraints

Integration & Operations Agents: 10. Integration Researcher - External dependencies and APIs 11. Database Researcher - Data modeling and query patterns 12. UI/UX Researcher - User interface patterns and best practices

Phase 3: Knowledge Integration

  1. Collect Research Findings

    • Gather insights from all research agents
    • Identify patterns and common themes
    • Evaluate conflicting recommendations
  2. Update Plan Sections

    • Enhance existing sections with new insights
    • Add new sections for uncovered aspects
    • Include references and citations
  3. Create Deepened Plan

    • Integrate research findings into coherent plan
    • Highlight key decisions and trade-offs
    • Provide implementation guidance

Phase 4: Validation and Refinement

  1. Review Deepened Plan

    • Check for consistency and completeness
    • Validate technical feasibility
    • Assess alignment with project goals
  2. Present Enhancements

    • Summarize key additions and improvements
    • Highlight critical insights and recommendations
    • Explain rationale for changes
  3. Next Steps

    • Option to further refine specific sections
    • Run adr docs/plans/<feature-name>.md — capture key architectural decisions as permanent ADRs
    • Start /core:work docs/plans/<feature-name>.md — begin implementation

Output Features

Enhanced Plan Sections:

  • Expanded technical specifications
  • Detailed implementation strategies
  • Comprehensive testing approaches
  • Performance and security considerations

Supporting Materials:

  • Reference links and documentation
  • Code examples and patterns
  • Architecture diagrams and flowcharts
  • Risk assessment and mitigation strategies

Key Benefits

  1. Depth: Uncover hidden complexities and edge cases
  2. Breadth: Consider multiple perspectives and approaches
  3. Validation: Ground decisions in research and best practices
  4. Actionability: Provide concrete implementation guidance

Use Cases

Best for:

  • Complex feature implementations
  • Architectural changes or migrations
  • Performance-critical optimizations
  • Security-sensitive functionality
  • Integration with external systems
Weekly Installs
6
First Seen
13 days ago
Installed on
augment6
github-copilot6
codex6
kimi-cli6
amp6
cline6