agentsclimarketplace

Subagent driven development

Skill charlieviettq/awesome-agent-skill/.cursor/skills/ai-agent-systems/subagent-driven-development

Curated skill pack for LLM agents in engineer and science workflow (Cursor & Claude ready).

Install
npx -y skills add charlieviettq/awesome-agent-skill --skill subagent-driven-development

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 22 stars22 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Execute an approved plan via one subagent per task with spec and quality checks. Use only when user allows subagents and a written plan exists. Triggers: "subagent per task", "dispatch implementer", "agent per plan step".

SKILL.md

2.0 KB, 404 tokens by cl100k_base, as published. Nobody here has run it

Subagent-driven development

Gate (required)

Use only if:

  • A written plan exists (planning-and-task-breakdown output) and user approved execution
  • User or parent explicitly allows subagents for implementation
  • Each task has clear files, acceptance, and verify steps

Otherwise use incremental-implementation in the main agent.

Workflow

For each plan task:
  1. Implementer subagent -> minimal change for task only
  2. Spec check -> matches acceptance criteria
  3. Quality check -> tests, style, scope (parent or review subagent)
  4. Parent integrates -> run verify-before-done for slice

Implementer brief (per task)

## Task
- **Title:**
- **Files (only these):**
- **Acceptance:**
- **Verify:**
- **Do not:** edit other files, refactor drive-by, commit unless instructed

Parent responsibilities

  • Own git strategy (commit per slice only if user/repo expects it)
  • Resolve conflicts between tasks
  • Stop plan if two tasks need the same file
  • Never claim full plan done until all tasks verified

Two-stage review

StageQuestion
Spec complianceDoes the change meet acceptance and stay in scope?
Code qualityTests, clarity, security, maintainability

Critical spec failures block the next task.

Anti-patterns

  • Subagents rewriting the plan mid-flight
  • Parallel implementers without dispatching-parallel-agents guards
  • Auto-commit/push from subagents without parent approval

Related

dispatching-parallel-agents, incremental-implementation, requesting-code-review, receiving-code-review, verify-before-done

Workflow inspired by obra/superpowers (MIT).

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.