subagent-driven-development

Agent workflows

obra/superpowersskills.sh ↗

Installs

194,717

deduplicated, at the last sync

Since we started

+1.9%

40 readings, about 2 hours apart. Not a live curve.

Our category

Agent workflows

ours

Last read

Sep 3, 2026

from the directory

Our brief

ours

This skill facilitates complex software development workflows by executing implementation plans in a structured, isolated manner. It dispatches specialized subagents for each task and performs detailed reviews after every step to ensure compliance with specifications and code quality. The process includes conflict resolution and final whole-branch review.

What it produces
  • Completed feature branch
  • Updated ledger file
What it needs
  • Implementation plan (file)
  • Spec document (optional)
Paid services

not detected

The evidence does not contain any information regarding paid services, APIs, or costs.

Registration

not detected

The evidence does not mention any registration requirements for using the skill.

Limits and human review

It cannot decide on conflicts or ambiguities without following its defined ruling process. It requires a plan to execute and must stop if it encounters an irreversible, destructive, security-sensitive action, or a broken plan where every path forward is a guess.

Evidenceskills/subagent-driven-development/SKILL.md:1-197
skills/subagent-driven-development/SKILL.md1–197 · excerpt truncated
---
name: subagent-driven-development
description: Use when executing implementation plans with independent tasks in the current session
---

# 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 the four named below, or all tasks complete. "Should I continue?" prompts and progress

Read 1 of 1 text files in the skill.

Checked September 2026 · GemmaWritten from the skill's published files. Check the source before relying on access or cost details.

Installfrom the directory

npx skills add https://github.com/obra/superpowers

Installing happens there, not here. We are an index with an opinion, not a mirror.

What is inside itfrom the directory

implementer-prompt.md
re-review-prompt.md
scripts/review-package
scripts/sdd-workspace
scripts/task-brief
SKILL.md
task-reviewer-prompt.md

7 files — names only. The directory does not report sizes.

What the auditors foundfrom the directory

A skill is instructions your agent will follow and scripts it may run, so who checked it matters as much as how many people installed it.

Gen Agent Trust HubThis skill implements a subagent-driven development workflow with specialized implementer and reviewer roles. It includes safety-oriented design patterns such as isolated workspaces and independent verification cycles. No malicious patterns or security vulnerabilities were detected.May 5, 2026 · SAFEpass
SocketNo alertsMay 5, 2026pass
SnykRisk: LOW · No issuesMay 5, 2026 · LOWpass
Runlayer4 files scanned · No issuesMar 12, 2026 · NONEpass
ZeroLeaksScore: 93/100 · 2 sections analyzedApr 16, 2026 · NONEpass

Installs, reading by readingours

191.9k
194.7k
Aug 31, 202640 readings over 4 days, drawn as the last reading of each day.Sep 3, 2026

Axis starts at 191.9k, not zero — the range is 191.9k to 194.7k.

Asked out loudspoken, not typed

The same skill in the words people use speaking to an assistant rather than typing into a box. Each one carries the situation it came from, and each answer says only what the skill's own files support.

asking about workflow structureIf I have multiple independent tasks in my plan, how does this work?

The skill executes each task by dispatching a fresh subagent. After every task, it performs a review to check for spec compliance and code quality. This process is designed for fast iteration without pausing between steps.

concerned about isolationDo I need to worry about context getting mixed up?

The skill uses fresh subagents per task, which means they operate with isolated context. This prevents them from inheriting or polluting the main session's history.

Filed alongside itours