subagent-driven-development

A guide for carrying out a written implementation plan through separate task workers with review checkpoints. Each worker handles an isolated task, followed by checks for requirement compliance and code quality.

In plain words
What is it for?
Use it when an existing plan can be divided into mostly independent tasks and the session supports isolated workers or a similar task-by-task workflow.
Why use it?
It helps coordinate multi-step work while keeping tasks and reviews distinct. This makes it easier to catch missed requirements or quality problems before considering the work complete.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/luizedupp/rememb/subagent-driven-development
Any agent
npx skills add LuizEduPP/Rememb --skill subagent-driven-development
Clone the repo
git clone --depth 1 https://github.com/LuizEduPP/Rememb

Made for: Claude Code, Codex.

Per session 34 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 834 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00034 $0.00834
Opus 5 $0.00017 $0.00417
Sonnet 5 $0.00007 $0.00167
Haiku 4.5 $0.00003 $0.00083

Measured yesterday against content hash e33f229c584e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

subagent-driven-development scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

src/rememb_skills/subagent-driven-development/SKILL.md · 117 lines

How it starts

The opening of the file, as written. The whole thing — 117 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Subagent-Driven Development

Overview

Use this skill to execute a written implementation plan through isolated workers with explicit review gates. If the current environment supports subagents, use them. If it does not, use the closest equivalent isolated-task workflow while keeping the same control pattern.

When to Use

Use this skill when:

  • a plan already exists
  • tasks are mostly independent
  • you want to preserve coordinator context while delegating implementation
  • you want spec compliance and code quality checked separately

Prefer a simpler inline workflow when tasks are tightly coupled or when the environment cannot isolate task execution meaningfully.

Core Workflow

  1. Read the implementation plan once and extract the full text for each task.
  2. Create a tracking list for all tasks before dispatching workers.
  3. For each task, dispatch one isolated implementer worker with the full task text and required context.
  4. If the implementer requests clarification, answer it before implementation continues.
  5. After implementation, run a spec-compliance review.
  6. Only after spec compliance passes, run a code-quality review.
  7. If either review finds issues, return the task to the implementer and repeat the relevant review.
  8. Mark the task complete only when both reviews pass.
  9. After all tasks are complete, run one final end-to-end review and verification pass.

Inputs / Assumptions

  • The plan is already written and broken into actionable tasks.
  • Each worker receives the task text directly rather than discovering it indirectly.
  • The coordinator keeps authority for sequencing, context, and approval.
  • Isolation matters more than raw speed for this workflow.

Worker Status Model

Implementers should report one of these statuses:

  • DONE: implementation completed and ready for review
  • DONE_WITH_CONCERNS: implementation completed, but there are correctness or scope concerns to inspect first
  • NEEDS_CONTEXT: missing information blocks safe completion
  • BLOCKED: the task cannot proceed under current assumptions or current worker capability

Read the full file on GitHub · 117 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 117 lines · 34 tokens per session scan A e33f229c584e

Subscribe to this mod's changes

subagent-driven-development is a skill published in the GitHub repository LuizEduPP/Rememb (4 stars, last pushed 1mo ago), licensed MIT. It adds 34 tokens to every session and 834 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

claude-api

Build, debug, and optimize Claude API / Anthropic SDK apps. Apps built with this skill should include prompt caching. Also handles migrating existing Claude API code between Claude model versions (4.5 → 4.6, 4.6 → 4.7, retired-model replacements). TRIGGER when: code imports anthropic/@anthropic-ai/sdk; user asks for…

Prismer-AI/PrismerCloud · 193 tokens

agent-coordination

Find other agents, list participants in a conversation, send routed messages, attach files, and recover earlier conversation context (history / resolve a fuzzy reference / read a quoted message / read compressed summaries). Use whenever you need to delegate to another agent, address a peer in a multi-agent…

Prismer-AI/PrismerCloud · 122 tokens

image-generate

Generate an image from a text prompt via the cloud LLM image proxy, persist it as a content-addressed workspace asset, and return a ContentBlock that downstream renderers can attach. Use whenever the user asks "draw / generate / make an image of …", an agent needs a diagram / illustration as a follow-up artifact, or a…

Prismer-AI/PrismerCloud · 103 tokens

claim-agent-ownership

Orchestrator skill for resolving multi-daemon binding contention. Use when you (the orchestrator) detect an agent.binding.contested sync event indicating two daemons are racing for the same agent — explicitly rebind ownership to a chosen target daemon so subsequent dispatches route deterministically. Implements Gap…

Prismer-AI/PrismerCloud · 101 tokens

prismer-im-collab

Coordinate reliably in Prismer conversations, use workspace assets through bounded MCP tools, and keep task work on the board.

Prismer-AI/PrismerCloud · 30 tokens

human-approval

Request human approval before performing a SAFETY-CRITICAL, IRREVERSIBLE, or SCOPE-EXPANDING action — submit a structured context (action, scope, risk, consequence) plus options, then STOP the current turn. The platform redispatches the agent after the human decides. NEVER use for routine deliverables (writing docs /…

Prismer-AI/PrismerCloud · 168 tokens