subagent-driven-development

Execute implementation plans by dispatching fresh implementer and reviewer subagents per task.

Updated Jul 30, 2026
One-click install
npx skills add https://github.com/johsquaree/claude-basic-docs --skill subagent-driven-development-johsquaree
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/johsquaree/claude-basic-docs/tree/main/.claude/skills/subagent-driven-development
Command: npx skills add https://github.com/johsquaree/claude-basic-docs --skill subagent-driven-development-johsquaree

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? Executing a multi-task implementation plan in one session pollutes context, loses progress after compaction, and ships unreviewed code. This Skill orchestrates plan execution through isolated subagents with per-task reviews and a persistent ledger. ## Core Features & Use Cases - Fresh subagent per task: Dispatch an implementer subagent with a task brief file, keeping the controller's context clean for coordination. - Two-stage review loop: Run a spec-compliance and code-quality review after every task, plus a final whole-branch review, with a bounded five-round fix loop and escalation to more capable models. - Crash-safe progress ledger: Track completed tasks, fix rounds, and parked findings in a per-plan workspace so work resumes correctly after context compaction. - Use Case: Given a written implementation plan with eight independent tasks, execute all of them in the current session — each implemented, tested, reviewed, and committed — then finish the development branch. ## Quick Start Execute the implementation plan at docs/plans/feature-plan.md using subagent-driven development, dispatching an implementer and reviewer for each task.

Frequently Asked Questions about subagent-driven-development

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I execute an implementation plan with subagents?

Dispatch a fresh implementer subagent per task with a task brief file, then dispatch a task reviewer on the resulting diff. Track progress in a ledger file and run a final whole-branch review after all tasks complete.

When should I use subagent-driven development vs executing plans directly?

Use it when you have an implementation plan with mostly independent tasks and want to stay in the current session. Tightly coupled tasks or plans needing a parallel session fit manual execution or a separate executing-plans workflow better.

How does the review fix loop work for each task?

Each task gets a spec-compliance and code-quality review. Findings trigger up to five fix rounds: rounds 1-3 resume the original implementer, rounds 4-5 dispatch a fresh implementer on a more capable model, each followed by a scoped re-review.

What happens if context is compacted mid-plan?

Progress survives in a per-plan ledger file under .superpowers/sdd/<plan-name>/, which records completed tasks, commits, and fix rounds. After compaction, the controller resumes from the ledger and git log rather than memory.

Which model should implementer and reviewer subagents use?

Always specify the model explicitly. Use cheap models for mechanical single-file tasks, standard models for multi-file integration, and the most capable model for architecture work and the final whole-branch review.