Subagent-Driven Development

Execute a plan by dispatching a fresh implementer subagent per task, a task review (spec compliance + code quality) after each task, and a broad whole-branch review at the end.

Why subagents: You delegate tasks to specialized agents with isolated context. By precisely crafting their instructions and context, you ensure they stay focused and succeed at their task. They should never inherit your session's context or history — you construct exactly what they need. This also preserves your own context for coordination work.

Core principle: Fresh subagent per task + task review (spec + quality) + broad final review = high quality, fast iteration

Narration: between tool calls, narrate at most one short line — the ledger and the tool results carry the record.

Pacing: Pacing is owned by the enclosing workflow. When run inside Sage's issue-implementation workflow, follow the selected HITL/AFK mode there; per-task commits always apply, while human gates may be deferred in AFK mode until final PR-readiness approval. Independent of pacing, these situations are escalation triggers — raise them with your human partner when they occur: a BLOCKED status you cannot resolve, ambiguity that genuinely prevents progress, or all tasks complete.

When to Use

Work through these questions in order:

1. Do you have a written implementation plan? If not, stop — write one first (use the writing-plans skill) or brainstorm the design before executing anything.
2. Are the tasks mostly independent? If they are tightly coupled — each task's design depends on discoveries made in the previous one — execute manually instead; per-task subagents would thrash.
3. Will execution happen in this session? If yes, use this skill. If the work must happen in a separate parallel session instead, this skill does not apply — coordinate that handoff outside it.