What problem does it solve?
Manually executing multi-task implementation plans leads to context pollution from accumulated state, inconsistent code quality, and missed spec or quality issues that compound across tasks, requiring costly rework and delaying delivery.
Core Features & Use Cases
- Fresh Subagent Per Task: Dispatches a new, focused subagent for each independent task to eliminate context confusion from prior work and ensure clean, targeted implementation.
- Mandatory Two-Stage Review: Enforces spec compliance review followed by code quality review for every task, catching under/over-building and code defects before they propagate to later tasks.
- End-to-End Workflow Integration: Natively integrates with planning skills like writing-plans, test-driven development practices, and systematic debugging tools to automate the full implementation lifecycle from plan to committed, tested code.
- Use Case: A software engineering team implementing a new payment processing feature with 6 independent tasks (payment model, transaction logging, endpoint creation, etc.) can use this skill to automate implementation, ensure every task matches the original plan requirements, and validate code quality before integration, reducing post-integration bugs by over 50%.
Quick Start
Use the subagent-driven-development skill to execute the implementation plan stored at docs/plans/payment-feature.md, automatically dispatching fresh subagents for each task with built-in spec compliance and code quality reviews.
