subagent-dev

subagent-dev is a skill for Claude Code, Codex from Insajin/autopus-adk. It costs 19 tokens per session (2,303 once invoked), scanned A, original, MIT.

A guide for designing and coordinating subagents, which are smaller coding agents delegated specific tasks. It covers their roles, permissions, tools, models, limits, memory, and isolated workspaces.

In plain words
What is it for?
Use it to define specialist agents for implementation, analysis, research, or background work.
Why use it?
It helps teams divide complex work safely and control what each delegated agent can do.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: mentions subagents.

Good fit Use it to define specialist agents for implementation, analysis, research, or background work.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/insajin/autopus-adk/subagent-dev
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.

Any agent
npx skills add Insajin/autopus-adk --skill subagent-dev
Clone the repo
git clone --depth 1 https://github.com/Insajin/autopus-adk

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for subagent-dev

README.md
[![agentmods](https://agentmods.dev/badge/skills/insajin/autopus-adk/subagent-dev/github.svg)](https://agentmods.dev/skills/insajin/autopus-adk/subagent-dev)
Your own site
<a href="https://agentmods.dev/skills/insajin/autopus-adk/subagent-dev"><img src="https://agentmods.dev/badge/skills/insajin/autopus-adk/subagent-dev/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for subagent-dev

Your own site · 80×15
<a href="https://agentmods.dev/skills/insajin/autopus-adk/subagent-dev"><img src="https://agentmods.dev/badge/skills/insajin/autopus-adk/subagent-dev.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 19 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,303 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
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 5.1 $0.00019 $0.02303
Opus 5 $0.00010 $0.01151
Sonnet 5 $0.00004 $0.00461
Haiku 4.5 $0.00002 $0.00230

Measured 8d ago against content hash 3b75dd7165c0, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

subagent-dev 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 8d ago.

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.

.omp/skills/subagent-dev/SKILL.md · 192 lines

How it starts

The opening of the file, as written. The whole thing — 192 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Subagent Development Skill

효과적인 서브에이전트를 설계하고 오케스트레이션하는 스킬입니다.

에이전트 정의 형식 (OMP 2.0+)

에이전트는 .omp/agents/<name>.md 에 YAML 프론트매터로 정의합니다.

프론트매터 필드

필드 필수 기본값 설명
name Yes - 고유 식별자, 소문자+하이픈
description Yes - 에이전트 역할 설명 (언제 위임할지 기준)
tools No 전체 상속 허용 도구 목록 (allowlist)
disallowedTools No 없음 차단 도구 목록 (denylist, tools와 상호배타)
model No 상속 opus, sonnet, haiku
permissionMode No default 권한 모드 (아래 표 참조)
maxTurns No 무제한 최대 에이전틱 턴 수
skills No 없음 에이전트 컨텍스트에 주입할 스킬 목록
hooks No 없음 에이전트 스코프 라이프사이클 훅
memory No 없음 영속 메모리 스코프 (user, project, local)
isolation No 없음 "worktree" — 격리된 git worktree에서 실행
background No false 백그라운드 실행 (대화 비차단)

권한 모드 (permissionMode)

모드 동작 적합한 에이전트
default 표준 권한 확인 범용 에이전트
acceptEdits 파일 편집 자동 승인 구현 에이전트
plan 읽기 전용, 쓰기 불가 분석/리서치 에이전트
dontAsk 모든 권한 요청 자동 거부 샌드박스 에이전트
bypassPermissions 모든 권한 검사 생략 완전 신뢰 자동화

정의 예시

---
name: expert-backend
description: 백엔드 API 구현 전문 에이전트. REST/gRPC 엔드포인트, 비즈니스 로직, DB 연동을 담당한다.
model: sonnet
tools: Read, Write, Edit, Grep, Glob, Bash
permissionMode: acceptEdits
maxTurns: 50
skills:
  - tdd
  - ddd
---

역할 지침:
1. 무엇을 해야 하는가
2. 어떤 파일을 수정할 수 있는가
3. 완료 기준은 무엇인가
4. 결과를 어떻게 보고하는가

서브에이전트 설계 원칙

단일 책임 원칙

각 에이전트는 하나의 명확한 역할을 가집니다:

  • 구현 에이전트: permissionMode: acceptEdits, tools에 Write/Edit 포함
  • 분석 에이전트: permissionMode: plan, 읽기 도구만 포함

격리 원칙

에이전트는 독립적으로 실행됩니다:

  • 이전 대화 히스토리에 접근 불가
  • 필요한 모든 컨텍스트를 spawn prompt에 포함
  • 결과는 구조화된 형식으로 반환
  • isolation: worktree — 여러 에이전트가 동시에 파일 수정 시 충돌 방지

병렬 실행 원칙

독립적인 작업은 병렬로 실행합니다:

{
  "i": "Dispatching bounded OMP work",
  "context": "Shared goal, constraints, owned-path boundaries, and cross-task contracts.",
  "tasks": [
    {
      "name": "Worker",
      "task": "Complete the assigned work and return the required receipt.",
      "outputSchema": {
        "type": "object",
        "additionalProperties": false,
        "required": ["owned_paths", "changed_files", "verification", "blockers", "next_required_step"],
        "properties": {
          "owned_paths": {"type": "array", "items": {"type": "string"}},
          "changed_files": {"type": "array", "items": {"type": "string"}},
          "verification": {"type": "array", "items": {"type": "string"}},
          "blockers": {"type": "array", "items": {"type": "string"}},
          "next_required_step": {"type": "string"}
        }
      },
      "schemaMode": "strict"
    }
  ]
}

Read the full file on GitHub · 192 lines

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. 8d ago First seen · 192 lines · 19 tokens per session scan A 3b75dd7165c0

Subscribe to this mod's changes

subagent-dev is a skill published in the GitHub repository Insajin/autopus-adk (111 stars, last pushed yesterday), licensed MIT. It adds 19 tokens to every session and 2,303 once invoked, about $0.0001 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 2026-09-03.