subagent-driven-developmentlisted

# Subagent-Driven Development Execute an implementation plan by giving each slice a focused implementation agent, then running two review gates before moving on. Core rule: the main session coordinates. Agents implement or review bounded work. The main session integrates, verifies, and decides when to proceed. ## When To Use Use this skill when: - An implementation plan already exists. - The plan has slices (`workbench:writing-plans` emits `### Slice N` sections) or similarly bounded tasks. - Slices can be assigned with clear ownership. - The current runtime has subagent support, or the main session can follow the same task discipline sequentially. If there is no plan yet, use `workbench:writing-plans` first. If implementation is starting, use `workbench:test-driven-development` before writing production code. ## Task Loop For each plan slice (or bounded task, when the plan is not slice-shaped): 1. Paste the slice text into the agent prompt: its heading, files map, signatures, test commands, acceptance criteria, and checkbox steps. One agent per slice; the checkbox steps are the slice's internal TDD rhythm and are never split across agents. 2. Define ownership: files, directories, or behavior the agent may change. 3. Include acceptance criteria and verification commands. 4. Tell the agent it is not alone in the codebase and must not revert edits made by others. 5. Wait for the implementation agent to return. 6. Record any deviations the agent returned in a running li