subagent-driven-development
Use when executing implementation plans with independent tasks in the current session
npx skills add guanyang/open-agent-hub --skill subagent-driven-development --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Execute plan by dispatching a fresh implementer subagent per task, a task review (spec compliance + code quality) after each, and a broad whole-branch review at the end.
How it works
- Dispatch an implementer subagent for each task (./implementer-prompt.md) and have the subagent ask questions when needed, then implement, test, commit, and perform self-reviews.
- Generate a review package with
scripts/review-package PLAN_FILE BASE HEADand dispatch a task reviewer (./task-reviewer-prompt.md). - Evaluate the task review for spec compliance and quality; if conflicts exist, address or escalate per the process, including rounds like: Fix round R of 5: R≤3 resume implementer; R≥4 fresh implementer, more capable model.
- After all findings are addressed, perform a final re-review (
./re-review-prompt.md) and then a final ledger update. - Maintain a ledger per plan in a dedicated workspace created via
scripts/sdd-workspace PLAN_FILE, store progress in<workspace>/progress.md, and ensure artifacts stay isolated to the plan’s workspace. - Use the plan’s scope to determine model selection and task complexity, ensuring each dispatch uses the appropriate model tier.
When to use it
- Use when there is an implementation plan with mostly independent tasks in the current session and you want to avoid context pollution by subagents.
- If staying in the same session with parallel task execution is appropriate, this approach contrasts with executing plans in a parallel session with shared context.
What it can touch
- Implementer subagent prompts at
./implementer-prompt.md. - Task review prompts at
./task-reviewer-prompt.md. - Re-review prompts at
./re-review-prompt.md. - The workspace setup and plan files are manipulated via
scripts/sdd-workspace PLAN_FILEand progress tracked in the plan’s<workspace>/progress.md.
Caveats
- Requires creating isolated workspaces per plan and strict ledger discipline to avoid cross-plan leakage.
- Back-and-forth rounds (R) may trigger more expensive models for complex rounds; resolution relies on the round rules embedded in the process.
- Some actions rely on human escalation if blockers cannot be resolved within the defined rules.
# Subagent-Driven Development Execute plan by dispatching a fresh implementer subagent per task, a task review (spec compliance + code quality) after each, 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. **Continuous execution:** Do not pause to check in with your human partner between tasks. Execute all tasks from the plan without stopping. The only reasons to stop are: BLOCKED status you cannot resolve, ambiguity that genuinely prevents progress, or all tasks complete. "Should I continue?" prompts and progress summaries waste their time — they asked you to execute the plan, so execute it. ## When
- When to Use
- The Process
- Setup
- Model Selection
- The Task Loop
- 1. Dispatch the implementer
- 2. Handle the report
- 3. Review the task
- 4. The fix loop
- 5. Complete the task
- Final Review
- Finish
- Common Rationalizations
- Example Workflow
What does the subagent-driven-development skill do?
Use when executing implementation plans with independent tasks in the current session
How do I install it?
Run `npx skills add guanyang/open-agent-hub --skill subagent-driven-development --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From guanyang/open-agent-hub, a repository with 940 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.