developing-with-subagents

Subagent-Driven Development

Contents

• Overview
• When to Use
• The Process
• Checkpoint Integration
• Related Skills

Overview

Execute plan by dispatching fresh subagent per task, with two-stage review after each: spec compliance review first, then code quality review.

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

Protocol Threshold (Required)

Follow the [CP Protocol Threshold] injected by hooks:

• Before the first Task call: output a standalone [CP1 Assessment] block (the block itself must not include tool calls; tool calls may follow in the same reply)
• Before claiming completion / requesting review / claiming verification passed: output a standalone [CP3 Assessment] block (the block itself must not include tool calls; the claim may follow in the same reply)

If unmet -> immediately perform the CP assessment, then continue the flow right away; do not stop or interrupt.

When to Use

digraph when_to_use {
    "Have implementation plan?" [shape=diamond];
    "Tasks mostly independent?" [shape=diamond];
    "Stay in this session?" [shape=diamond];
    "developing-with-subagents" [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?" -> "developing-with-subagents" [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)
• Two-stage review after each task: spec compliance first, then code quality
• Faster iteration (no human-in-loop between tasks)

The Process

Copy this checklist template to track overall progress:

Plan Execution Progress:
- [ ] Plan read and all tasks extracted
- [ ] TodoWrite created with all tasks

Per-Task Checklist (copy for each):
Task N: [description]
- [ ] Checkpoint 1 (Task Analysis) applied
- [ ] 【CP1 Assessment】 output (standalone message)
- [ ] Implementer subagent dispatched
- [ ] Questions answered (if any)
- [ ] Implementation complete
- [ ] Checkpoint 3 (Quality Gate) applied
- [ ] Spec reviewer: ✅ compliant
- [ ] Code quality reviewer: ✅ approved
- [ ] Task marked complete

Final Steps:
- [ ] All tasks complete
- [ ] Final code reviewer dispatched
- [ ] finishing-development-branches invoked

digraph process {
    rankdir=TB;

    subgraph cluster_per_task {
        label="Per Task";
        "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 (./code-quality-reviewer-prompt.md)" [shape=box];
        "Code quality reviewer subagent approves?" [shape=diamond];
        "Implementer subagent fixes quality issues" [shape=box];
        "Mark task complete in TodoWrite" [shape=box];
    }

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

    "Read plan, extract all tasks with full text, note context, create TodoWrite" -> "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 (./code-quality-reviewer-prompt.md)" [label="yes"];
    "Dispatch code quality reviewer subagent (./code-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 (./code-quality-reviewer-prompt.md)" [label="re-review"];
    "Code quality reviewer subagent approves?" -> "Mark task complete in TodoWrite" [label="yes"];
    "Mark task complete in TodoWrite" -> "More tasks remain?";
    "More tasks remain?" -> "Dispatch implementer subagent (./implementer-prompt.md)" [label="yes"];
    "More tasks remain?" -> "Dispatch final code reviewer subagent for entire implementation" [label="no"];
    "Dispatch final code reviewer subagent for entire implementation" -> "Use superpowers:finishing-development-branches";
}

Prompt Templates

• ./implementer-prompt.md - Dispatch implementer subagent
• ./spec-reviewer-prompt.md - Dispatch spec compliance reviewer subagent
• ./code-quality-reviewer-prompt.md - Dispatch code quality reviewer subagent

Claude Model Strategy

Choose model based on task type:

Subagent	Model	Freedom
Implementer	model: sonnet	Low - always use Sonnet for code
Spec/Quality Reviewer	Opus (default)	Low - always use Opus for review
Exploration	model: haiku	Medium - prefer Haiku, flexible

Collaboration Checkpoints

Apply checkpoint logic from coordinating-multi-model-work/checkpoints.md at these stages:

► Checkpoint 1 (Task Analysis): Before dispatching implementer subagent:

• Collect: task files, description, complexity
• Output a standalone 【CP1 Assessment】 block to the user BEFORE the first Task tool call
• Check critical task conditions → Match: invoke expert model
• Evaluate general task signals → Positive: invoke

► Checkpoint 2 (Mid-Review): During subagent execution:

• Subagent asks question requiring external expertise → invoke domain expert
• Multiple implementation approaches debated → invoke cross-validation

► Checkpoint 3 (Quality Gate): After subagent completes, before spec review:

• Implementation complete → invoke domain expert for pre-review assessment

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 TodoWrite with all tasks]

Task 1: Hook installation script

[Get Task 1 text and context (already extracted)]
[Dispatch implementation subagent with full task text + context]

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

[Dispatch spec compliance reviewer]
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.

[Mark Task 1 complete]

Task 2: Recovery modes

[Get Task 2 text and context (already extracted)]
[Dispatch implementation subagent with full task text + context]

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

[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)

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)

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

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 (spec compliance OR code quality)
• Proceed with unfixed issues
• 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 (spec reviewer found issues = not done)
• Skip review loops (reviewer found issues = implementer fixes = review again)
• Let implementer self-review replace actual review (both are needed)
• Start code quality review before spec compliance is ✅ (wrong order)
• Move to next task while either review has open issues

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:

• superpowers:writing-plans - Creates the plan this skill executes
• superpowers:requesting-code-review - Code review template for reviewer subagents
• superpowers:finishing-development-branches - Complete development after all tasks
• superpowers:coordinating-multi-model-work - Multi-model routing for task execution

Subagents should use:

• superpowers:practicing-test-driven-development - Subagents follow TDD for each task

Alternative workflow:

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

Multi-Model Task Dispatch

Related skill: superpowers:coordinating-multi-model-work

At checkpoints, apply semantic routing from coordinating-multi-model-work/routing-decision.md:

• Routing decision:

  • Clear backend task (API, database, server logic) → CODEX
  • Clear frontend task (UI, components, styles) → GEMINI
  • Full-stack integration or critical task → CROSS_VALIDATION
  • General task requiring complete context → CLAUDE subagent

• Check for Model Hint: If task includes hint, use as guidance

• Notify user: "I will use [model/subagent] to implement [task name]"

• Call MCP tool with English prompts (see coordinating-multi-model-work/INTEGRATION.md for templates). Use Codex MCP (mcp__codex__codex) for backend, Gemini MCP (mcp__gemini__gemini) for frontend, and call both in parallel for CROSS_VALIDATION.

Full checkpoint logic: See coordinating-multi-model-work/checkpoints.md

Fallback (Fail-Closed): If external models are required but unavailable or time out, STOP and follow coordinating-multi-model-work/GATE.md (do not proceed with task completion output).