What problem does it solve? Subagents do not inherit parent context, so vague delegation prompts force them to rediscover the codebase, waste time, and risk conflicting edits. This Skill ensures every subagent spawn call includes concrete file paths, architecture context, ownership boundaries, and expected outputs. ## Core Features & Use Cases - Mandatory Brief Checklist: Enforces goal, starting files, architecture context, scope boundary, ownership, related files, expected output, and a non-reversion coordination rule in every subagent prompt. - Explorer and Worker Templates: Provides distinct briefing patterns for read-only research agents (requiring file:line evidence) and implementation agents (requiring disjoint write ownership). - Copy-Ready Task Templates: Ships references/task-templates.md with fill-in-the-blank prompts for bug fixes, new features, research, refactors, and multi-file edits. - Use Case: Before spawning parallel workers to implement disjoint modules, generate worker briefs that assign owned file globs, list read-only related files, and forbid reverting peer changes. ## Quick Start Ask the assistant to brief an explorer subagent to map how a specific module works, providing the starting file paths and the questions to answer.
