subagent-driven（子 Agent 驱动）

Subagent-Driven Development

Execute plan by dispatching fresh subagent per task, with two-stage review after each: spec compliance review first, then code quality review. Each subagent gets its own planning directory for structured knowledge capture.

Core principle: Fresh subagent per task + per-agent planning dir + two-stage review (spec then quality) = high quality, fast iteration

NON-NEGOTIABLE: Two-Stage Review Gate

1. Spec Compliance Review — Dispatch ./spec-reviewer-prompt.md subagent
2. Code Quality Review — Dispatch ./quality-reviewer-prompt.md subagent (only after spec review passes)

A task is NOT complete until BOTH reviews return APPROVED. No exceptions — not for "simple" tasks, config changes, or thorough self-reviews.

The Task Status Dashboard in .planning/progress.md has Spec Review and Quality Review columns. Both MUST show PASS before status can be complete.

When to Use

digraph when_to_use {
    "Have implementation plan?" [shape=diamond];
    "Tasks mostly independent?" [shape=diamond];
    "Stay in this session?" [shape=diamond];
    "subagent-driven" [shape=box];
    "executing-plans" [shape=box];
    "Manual execution or brainstorm first" [shape=box];

    "Have implementation plan?" -> "Tasks mostly independent?" [label="yes"];
    "Have implementation plan?" -> "Manual execution or brainstorm first" [label="no"];
    "Tasks mostly independent?" -> "Stay in this session?" [label="yes"];
    "Tasks mostly independent?" -> "Manual execution or brainstorm first" [label="no - tightly coupled"];
    "Stay in this session?" -> "subagent-driven" [label="yes"];
    "Stay in this session?" -> "executing-plans" [label="no - parallel session"];
}

vs. Executing Plans (parallel session):

• Same session (no context switch)
• Fresh subagent per task (no context pollution)
• Per-agent planning directories (structured knowledge capture)
• Two-stage review after each task: spec compliance first, then code quality
• Faster iteration (no human-in-loop between tasks)

The Process

digraph process {
    rankdir=TB;

    subgraph cluster_per_task {
        label="Per Task";
        "Create agent planning dir (if not exists)" [shape=box style=filled fillcolor=lightyellow];
        "Dispatch implementer subagent (./implementer-prompt.md)" [shape=box];
        "Implementer subagent asks questions?" [shape=diamond];
        "Answer questions, provide context" [shape=box];
        "Implementer subagent implements, tests, commits, self-reviews" [shape=box];
        "Dispatch spec reviewer subagent (./spec-reviewer-prompt.md)" [shape=box];
        "Spec reviewer subagent confirms code matches spec?" [shape=diamond];
        "Implementer subagent fixes spec gaps" [shape=box];
        "Dispatch code quality reviewer subagent (./quality-reviewer-prompt.md)" [shape=box];
        "Code quality reviewer subagent approves?" [shape=diamond];
        "Implementer subagent fixes quality issues" [shape=box];
        "Aggregate agent findings into top-level .planning/" [shape=box style=filled fillcolor=lightyellow];
        "Mark task complete via TaskUpdate" [shape=box];
    }

    "Read plan, extract all tasks with full text, note context, create tasks via TaskCreate" [shape=box];
    "More tasks remain?" [shape=diamond];
    "Dispatch final code reviewer subagent for entire implementation" [shape=box];
    "Use superpower-planning:finishing-branch" [shape=box style=filled fillcolor=lightgreen];

    "Read plan, extract all tasks with full text, note context, create tasks via TaskCreate" -> "Create agent planning dir (if not exists)";
    "Create agent planning dir (if not exists)" -> "Dispatch implementer subagent (./implementer-prompt.md)";
    "Dispatch implementer subagent (./implementer-prompt.md)" -> "Implementer subagent asks questions?";
    "Implementer subagent asks questions?" -> "Answer questions, provide context" [label="yes"];
    "Answer questions, provide context" -> "Dispatch implementer subagent (./implementer-prompt.md)";
    "Implementer subagent asks questions?" -> "Implementer subagent implements, tests, commits, self-reviews" [label="no"];
    "Implementer subagent implements, tests, commits, self-reviews" -> "Dispatch spec reviewer subagent (./spec-reviewer-prompt.md)";
    "Dispatch spec reviewer subagent (./spec-reviewer-prompt.md)" -> "Spec reviewer subagent confirms code matches spec?";
    "Spec reviewer subagent confirms code matches spec?" -> "Implementer subagent fixes spec gaps" [label="no"];
    "Implementer subagent fixes spec gaps" -> "Dispatch spec reviewer subagent (./spec-reviewer-prompt.md)" [label="re-review"];
    "Spec reviewer subagent confirms code matches spec?" -> "Dispatch code quality reviewer subagent (./quality-reviewer-prompt.md)" [label="yes"];
    "Dispatch code quality reviewer subagent (./quality-reviewer-prompt.md)" -> "Code quality reviewer subagent approves?";
    "Code quality reviewer subagent approves?" -> "Implementer subagent fixes quality issues" [label="no"];
    "Implementer subagent fixes quality issues" -> "Dispatch code quality reviewer subagent (./quality-reviewer-prompt.md)" [label="re-review"];
    "Code quality reviewer subagent approves?" -> "Aggregate agent findings into top-level .planning/" [label="yes"];
    "Aggregate agent findings into top-level .planning/" -> "Mark task complete via TaskUpdate";
    "Mark task complete via TaskUpdate" -> "More tasks remain?";
    "More tasks remain?" -> "Create agent planning dir (if not exists)" [label="yes - next task"];
    "More tasks remain?" -> "Dispatch final code reviewer subagent for entire implementation" [label="no"];
    "Dispatch final code reviewer subagent for entire implementation" -> "Use superpower-planning:finishing-branch";
}

Per-Agent Planning Directories

Each agent role gets ONE directory, reused across all tasks:

mkdir -p .planning/agents/{role}/

Example:

mkdir -p .planning/agents/implementer/
mkdir -p .planning/agents/spec-reviewer/
mkdir -p .planning/agents/quality-reviewer/

Each agent planning dir contains:

• findings.md - Discoveries, decisions, critical items (appended across tasks)
• progress.md - Step-by-step progress log (appended across tasks)

Do NOT create per-task directories like implementer-task-1/, implementer-task-2/. One directory per role, updated continuously.

Include the planning dir path in the agent's prompt using ./implementer-prompt.md template.

Orchestrator Aggregation Flow

After each task completes (both reviews passed), aggregate the agent's findings:

1. Read the agent's .planning/agents/{role}/findings.md and progress.md
2. Extract items marked with > **Critical for Orchestrator:** plus any errors and test results
3. Append extracted items to top-level .planning/findings.md under a task heading
4. Update top-level .planning/progress.md:

- Update the Task Status Dashboard table at the top (add/update the row for this task) - Append completion details to the session log section

Example aggregation:

<!-- Append to .planning/findings.md -->
## Task 2: Recovery modes
- [From implementer] Database migration requires careful ordering
- [From spec-reviewer] All requirements met after fix pass
- [From quality-reviewer] Approved with no issues

<!-- Update Task Status Dashboard table in .planning/progress.md -->
| Task 1: Hook installation | ✅ complete | PASS | PASS | agents/implementer/ | 5 tests passing |
| Task 2: Recovery modes | ✅ complete | PASS (2nd pass) | PASS | agents/implementer/ | 8 tests passing |
| Task 3: Config parser | ⏳ pending | - | - | - | - |

<!-- Append to session log in .planning/progress.md -->
- [x] Task 2: Recovery modes - COMPLETED
  - Implementer: 8 tests passing, committed
  - Spec review: Passed (2nd pass after fix)
  - Quality review: Approved

Prompt Templates

• ./implementer-prompt.md - Dispatch implementer subagent (includes planning dir injection)
• ./spec-reviewer-prompt.md - Dispatch spec compliance reviewer subagent
• ./quality-reviewer-prompt.md - Dispatch code quality reviewer subagent

Example Workflow

You: I'm using Subagent-Driven Development to execute this plan.

[Read plan file once: docs/plans/feature-plan.md]
[Extract all 5 tasks with full text and context]
[Create all tasks via TaskCreate]

Task 1: Hook installation script

[Create .planning/agents/implementer/ (if not exists)]
[Dispatch implementation subagent with full task text + context + planning dir]

Implementer: "Before I begin - should the hook be installed at user or system level?"

You: "User level (~/.config/superpowers/hooks/)"

Implementer: "Got it. Implementing now..."
[Later] Implementer:
  - Implemented install-hook command
  - Added tests, 5/5 passing
  - Self-review: Found I missed --force flag, added it
  - Committed
  - Logged findings to .planning/agents/implementer/findings.md

[Dispatch spec compliance reviewer with its own planning dir]
Spec reviewer: Spec compliant - all requirements met, nothing extra

[Get git SHAs, dispatch code quality reviewer]
Code reviewer: Strengths: Good test coverage, clean. Issues: None. Approved.

[Aggregate: read agent findings, append to .planning/findings.md and progress.md]
[Mark Task 1 complete]

Task 2: Recovery modes

[Reuse .planning/agents/implementer/ (already exists from Task 1)]
[Dispatch implementation subagent with full task text + context + planning dir]

Implementer: [No questions, proceeds]
Implementer:
  - Added verify/repair modes
  - 8/8 tests passing
  - Self-review: All good
  - Committed

[Dispatch spec compliance reviewer]
Spec reviewer: Issues:
  - Missing: Progress reporting (spec says "report every 100 items")
  - Extra: Added --json flag (not requested)

[Implementer fixes issues]
Implementer: Removed --json flag, added progress reporting

[Spec reviewer reviews again]
Spec reviewer: Spec compliant now

[Dispatch code quality reviewer]
Code reviewer: Strengths: Solid. Issues (Important): Magic number (100)

[Implementer fixes]
Implementer: Extracted PROGRESS_INTERVAL constant

[Code reviewer reviews again]
Code reviewer: Approved

[Aggregate agent findings into .planning/]
[Mark Task 2 complete]

...

[After all tasks]
[Dispatch final code-reviewer]
Final reviewer: All requirements met, ready to merge

Done!

Advantages

vs. Manual execution:

• Subagents follow TDD naturally
• Fresh context per task (no confusion)
• Parallel-safe (subagents don't interfere)
• Subagent can ask questions (before AND during work)
• Per-agent planning dirs capture knowledge persistently

vs. Executing Plans:

• Same session (no handoff)
• Continuous progress (no waiting)
• Review checkpoints automatic

Efficiency gains:

• No file reading overhead (controller provides full text)
• Controller curates exactly what context is needed
• Subagent gets complete information upfront
• Questions surfaced before work begins (not after)
• Planning dirs prevent knowledge loss between subagents

Quality gates:

• Self-review catches issues before handoff
• Two-stage review: spec compliance, then code quality
• Review loops ensure fixes actually work
• Spec compliance prevents over/under-building
• Code quality ensures implementation is well-built
• Aggregation preserves findings for future tasks

Cost:

• More subagent invocations (implementer + 2 reviewers per task)
• Controller does more prep work (extracting all tasks upfront)
• Review loops add iterations
• But catches issues early (cheaper than debugging later)

Red Flags

Never:

• Skip reviews — see Two-Stage Review Gate above. No exceptions.
• Start implementation on main/master branch without explicit user consent
• Dispatch multiple implementation subagents in parallel (conflicts)
• Make subagent read plan file (provide full text instead)
• Skip scene-setting context (subagent needs to understand where task fits)
• Ignore subagent questions (answer before letting them proceed)
• Accept "close enough" on spec compliance (reviewer found issues = not done)
• Start code quality review before spec compliance passes (wrong order)
• Skip planning dir creation or aggregation step (knowledge gets lost)

If subagent asks questions:

• Answer clearly and completely
• Provide additional context if needed
• Don't rush them into implementation

If reviewer finds issues:

• Implementer (same subagent) fixes them
• Reviewer reviews again
• Repeat until approved
• Don't skip the re-review

If subagent fails task:

• Dispatch fix subagent with specific instructions
• Don't try to fix manually (context pollution)

Integration

Required workflow skills:

• superpower-planning:git-worktrees - RECOMMENDED: Set up isolated workspace unless already on a feature branch
• superpower-planning:writing-plans - Creates the plan this skill executes
• superpower-planning:requesting-review - Code review template for reviewer subagents
• superpower-planning:finishing-branch - Complete development after all tasks

Subagents should use:

• superpower-planning:tdd - Subagents follow TDD for each task

Alternative workflow:

• superpower-planning:executing-plans - Use for parallel session instead of same-session execution