subagent-driven-development

Dispatching independent subagents for each task isolates context and enables fast iteration. This approach prevents context leakage by creating a dedicated implementer for every task, ensuring structured execution of complex implementation plans.

Two-stage review first verifies spec compliance, then assesses code quality before task completion. You trigger the spec compliance reviewer after a task completes, followed by the code quality reviewer, iterating until both pass before proceeding.

Yes, subagent task dispatch suits complex implementation workflows where tasks are largely autonomous. It applies to scenarios requiring per-task delegation, context isolation, and controlled handoffs via predefined prompts and templates.

Per-task subagent dispatch prevents context leakage by creating a dedicated implementer subagent for every task. This isolates the context and ensures fast iteration through structured two-stage reviews.

Yes, predefined prompt templates are required for controlled handoffs. They provide the full task text and context needed to dispatch implementer subagents and trigger the staged reviewers for spec compliance and code quality.

Avoid subagent-driven development when tasks lack autonomy or require continuous shared context. This skill satisfies requirements for largely autonomous tasks needing context isolation, staged reviews, and controlled handoffs via predefined templates.