agentleFS
Tsuchiya2Tsuchiya2/evaluator-driven-agent-flow/.claude/skills/edaf-orchestration/SKILL.md

edaf-orchestration

Skill1 starsChanged 10 months agoMITView on GitHub

Skill Type: Workflow Orchestration Used By: Main Agent Trigger: When user requests feature implementation with "エージェントフロー" or "EDAF" This skill provides the 7-Phase Gate System workflow for EDAF (Evaluator-Driven Agent Flow). It defines the complete workflow from requirements gathering to deployment with quality gates at each phase. Use this skill when user says: - "エージェントフローで実装して" - "EDAFで開発して" - "Use EDAF for this feature" - "Implement with agent flow"

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Trigger: When user requests feature implementation with "エージェントフロー" or "EDAF"

    line 5
  2. Distinctive:

    Use this skill when user says:

    line 17

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

EDAF Orchestration Skill

Skill Type: Workflow Orchestration Used By: Main Agent Trigger: When user requests feature implementation with "エージェントフロー" or "EDAF"


Overview

This skill provides the 7-Phase Gate System workflow for EDAF (Evaluator-Driven Agent Flow). It defines the complete workflow from requirements gathering to deployment with quality gates at each phase.


Trigger Conditions

Use this skill when user says:

  • "エージェントフローで実装して"
  • "EDAFで開発して"
  • "Use EDAF for this feature"
  • "Implement with agent flow"

Skill Contents

FilePurpose
SKILL.mdThis overview file
PHASE1-REQUIREMENTS.mdPhase 1: Requirements Gathering Gate workflow
PHASE2-DESIGN.mdPhase 2: Design Gate workflow
PHASE3-PLANNING.mdPhase 3: Planning Gate workflow
PHASE4-IMPLEMENTATION.mdPhase 4: Implementation workflow with quality gate
PHASE5-CODE.mdPhase 5: Code Review Gate workflow
PHASE6-DOCUMENTATION.mdPhase 6: Documentation Update workflow
PHASE7-DEPLOYMENT.mdPhase 7: Deployment Gate workflow (optional)
GATE-PATTERNS.mdCommon gate patterns and pass criteria

Quick Reference

Phase Overview

Phase 1: Requirements Gathering Gate
├── Requirements Gatherer (interactive 5W2H dialogue)
└── 7 Requirements Evaluators approve (≥8.0/10.0)
    ↓
Phase 2: Design Gate
├── Designer creates design document
└── 7 Design Evaluators approve (≥8.0/10.0)
    ↓
Phase 3: Planning Gate
├── Planner creates task plan
└── 7 Planner Evaluators approve (≥8.0/10.0)
    ↓
Phase 4: Implementation
├── Database Worker
├── Backend Worker
├── Frontend Worker
├── Test Worker
└── Quality Gate Evaluator (Score 10.0 = PASS: lint + tests)
    ↓
Phase 5: Code Review Gate
├── 7 Code Evaluators approve (≥8.0/10.0)
└── UI Verification (if frontend changed)
    ↓
Phase 6: Documentation Update
├── Documentation Worker updates permanent docs
└── 5 Documentation Evaluators approve (≥8.0/10.0)
    ↓
Phase 7: Deployment Gate (Optional)
└── 5 Deployment Evaluators approve (≥8.0/10.0)

Critical Rules

  1. NEVER skip phases
  2. ALWAYS launch evaluators in parallel
  3. WAIT for ALL evaluators to approve before proceeding
  4. Pass criteria:
    • Phases 1, 2, 3, 5, 6, 7: ≥8.0/10.0
    • Phase 4 Quality Gate: 10.0/10.0 (zero errors, zero warnings, all tests passing)

Phase Flow Control

Starting EDAF

// Update status
await bash('.claude/scripts/update-edaf-phase.sh "Starting" "Initializing EDAF workflow"')

// Begin Phase 1
// Reference: PHASE1-DESIGN.md

Phase Transition

// After Phase 1 passes
await bash('.claude/scripts/update-edaf-phase.sh "Phase 1" "Complete - 7/7 evaluators passed"')

// Transition to Phase 2
await bash('.claude/scripts/update-edaf-phase.sh "Phase 2" "Starting planning"')

Evaluator Reference

Phase 1: Requirements Evaluators (7)

  • requirements-clarity-evaluator
  • requirements-completeness-evaluator
  • requirements-feasibility-evaluator
  • requirements-goal-alignment-evaluator
  • requirements-scope-evaluator
  • requirements-testability-evaluator
  • requirements-user-value-evaluator

Phase 2: Design Evaluators (7)

  • design-consistency-evaluator
  • design-extensibility-evaluator
  • design-goal-alignment-evaluator
  • design-maintainability-evaluator
  • design-observability-evaluator
  • design-reliability-evaluator
  • design-reusability-evaluator

Phase 3: Planner Evaluators (7)

  • planner-clarity-evaluator
  • planner-deliverable-structure-evaluator
  • planner-dependency-evaluator
  • planner-goal-alignment-evaluator
  • planner-granularity-evaluator
  • planner-responsibility-alignment-evaluator
  • planner-reusability-evaluator

Phase 4: Quality Gate Evaluator (1)

  • quality-gate-evaluator (checks both lint and tests, requires 10.0/10.0)

Phase 5: Code Evaluators (7)

  • code-quality-evaluator-v1-self-adapting
  • code-testing-evaluator-v1-self-adapting
  • code-security-evaluator-v1-self-adapting
  • code-documentation-evaluator-v1-self-adapting
  • code-maintainability-evaluator-v1-self-adapting
  • code-performance-evaluator-v1-self-adapting
  • code-implementation-alignment-evaluator-v1-self-adapting

Phase 6: Documentation Evaluators (5)

  • documentation-completeness-evaluator
  • documentation-accuracy-evaluator
  • documentation-consistency-evaluator
  • documentation-clarity-evaluator
  • documentation-currency-evaluator

Phase 7: Deployment Evaluators (5)

  • deployment-readiness-evaluator
  • production-security-evaluator
  • observability-evaluator
  • performance-benchmark-evaluator
  • rollback-plan-evaluator

Related Resources

  • Workers: .claude/agents/workers/
  • Evaluators: .claude/agents/evaluators/
  • Scoring Skill: .claude/skills/edaf-evaluation/
  • UI Verification Skill: .claude/skills/ui-verification/

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.