Epic Walkthrough Skill

Generate comprehensive walkthrough documentation and trigger self-improvement for epic workflows.

Purpose

Create walkthrough.md documenting the epic journey and trigger workflow audit/healing for continuous improvement.

When to Invoke

• During

  /epic

  Step 8 (Finalization phase)
• After deployment completes successfully
• Via

  /finalize

  command

Template Reference

Use the template at

.spec-flow/templates/epic/walkthrough.md

as the base structure.

Template variables (replace with actual values):

• {{EPIC_TITLE}}

  ,

  {{EPIC_NUMBER}}

  ,

  {{EPIC_SLUG}}

• {{START_DATE}}

  ,

  {{END_DATE}}

  ,

  {{DURATION_DAYS}}

  ,

  {{DURATION_HOURS}}

• {{SPRINT_COUNT}}

  ,

  {{TASKS_COMPLETED}}

• {{EXPECTED_VELOCITY}}

  ,

  {{ACTUAL_VELOCITY}}

• Phase durations, sprint details, quality gate results
• File change statistics, deployment information

Data sources:

• state.yaml

  - Phase timestamps, sprint status

• epic-spec.md

  - Business value, constraints

• sprint-plan.md

  - Sprint dependencies, estimates

• optimization-report.md

  - Quality gate results

• git log

  - Commit history, file statistics

Walkthrough Structure

<walkthrough>
  <overview>
    <epic_goal>...</epic_goal>
    <business_value>...</business_value>
    <duration>Start: 2025-11-19, End: 2025-11-25 (6 days)</duration>
    <velocity>Expected: 3x, Actual: 4.2x</velocity>
  </overview>

  <phases_completed>
    <phase name="research">
      <duration>2 hours</duration>
      <key_findings>OAuth 2.1 recommended, JWT tokens for sessions</key_findings>
      <artifacts>research.md, API contracts locked</artifacts>
    </phase>
    <phase name="planning">
      <duration>1.5 hours</duration>
      <key_decisions>PostgreSQL for sessions, httpOnly cookies</key_decisions>
      <artifacts>plan.md, sprint-plan.md</artifacts>
    </phase>
    <phase name="implementation">
      <duration>16 hours</duration>
      <sprints_completed>S01 (8h), S02 (12h), S03 (6h)</sprints_completed>
      <artifacts>Implementation in sprints/S01, S02, S03</artifacts>
    </phase>
    <phase name="optimization">
      <duration>2 hours</duration>
      <quality_gates>All passed (tests, security, performance, accessibility)</quality_gates>
      <artifacts>optimization-report.xml</artifacts>
    </phase>
  </phases_completed>

  <validation>
    <quality_gates>
      <gate name="tests" status="passed">98% coverage, 142/142 passing</gate>
      <gate name="security" status="passed">0 critical, 2 medium (triaged)</gate>
      <gate name="performance" status="passed">API p95: 87ms (target: <200ms)</gate>
      <gate name="accessibility" status="passed">WCAG 2.1 AA compliant</gate>
    </quality_gates>
    <deployment>
      <staging>Deployed v1.5.0-rc.1, validated 2025-11-24</staging>
      <production>Promoted to v1.5.0, deployed 2025-11-25</production>
    </deployment>
  </validation>

  <key_files_modified>
    <category name="backend">
      <file path="src/auth/middleware.ts">JWT validation middleware</file>
      <file path="src/auth/service.ts">OAuth 2.1 integration</file>
    </category>
    <category name="frontend">
      <file path="components/LoginForm.tsx">OAuth login flow</file>
      <file path="hooks/useAuth.ts">Auth state management</file>
    </category>
    <category name="infrastructure">
      <file path="migrations/20251119_sessions.sql">Session table</file>
      <file path="contracts/api/auth-v1.yaml">Auth API contract</file>
    </category>
  </key_files_modified>

  <next_steps>
    <future_enhancement priority="high">Add MFA support</future_enhancement>
    <future_enhancement priority="medium">Social login</future_enhancement>
    <technical_debt>Token refresh uses polling (should use refresh tokens)</technical_debt>
    <monitoring>Watch auth_failures_total metric</monitoring>
  </next_steps>

  <summary>
    <what_worked>
      - Parallel sprint execution saved 8 hours
      - API contract locking prevented integration bugs
      - Meta-prompting research identified OAuth 2.1 early
    </what_worked>
    <what_struggled>
      - Sprint S02 took 12h vs 8h estimated
      - Cookie SameSite policy required research
    </what_struggled>
    <lessons_learned>
      - Lock contracts before parallel work
      - Research phase crucial for architectural decisions
      - Parallel execution 4.2x faster than sequential
    </lessons_learned>
    <velocity_impact>
      Expected: 3x faster
      Actual: 4.2x faster
      Reason: S02/S03 ran 60% parallel
    </velocity_impact>
  </summary>
</walkthrough>

Self-Improvement Actions

1. Run post-mortem audit:

/audit-workflow

2. Pattern detection (after 2-3 epics):

• Analyze completed epics for recurring patterns
• Detect code generation opportunities
• Suggest custom skills for detected patterns
• Generate project-specific commands/hooks

3. Self-healing:

/heal-workflow

• Apply approved workflow improvements
• Update skills based on learnings

4. Update documentation:

• CHANGELOG.md
• README.md
• Project CLAUDE.md (add epic summary)

5. Archive epic artifacts:

• Move prompts to completed/
• Mark epic complete in state.yaml

Outputs

• walkthrough.md

  - Comprehensive epic summary

• workflow-improvements.xml

  - Self-improvement suggestions
• Updated project documentation

State Updates

After finalization:

phases:
  finalization:
    status: completed
    completed_at: { ISO_TIMESTAMP }

epic:
  status: completed
  completed_at: { ISO_TIMESTAMP }
  velocity_actual: 4.2

Velocity Tracking

Metrics captured:

• Expected vs actual sprint durations
• Parallel execution efficiency
• Quality gate pass rates
• Deployment success rate

Used for:

• Future epic estimation
• Workflow optimization
• Pattern detection for self-improvement