A development-control skill for entwurf, a software project with multiple agent sessions called garden citizens. It lets an agent inspect, open, message, resume, and test those sessions.

Use it to list sessions, open Pi or Claude Code workers, send messages, resume dormant sessions, and inspect the system boundary.

It gives developers a way to manage separate coding-agent sessions and connect replies to the correct session.

How junghan0611/entwurf configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to junghan0611/entwurf. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/junghan0611/entwurf/main/.claude/skills/entwurf-dev/SKILL.md

Clone the repo

git clone --depth 1 https://github.com/junghan0611/entwurf

Made for: Claude Code.

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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/junghan0611/entwurf/entwurf-dev.svg)](https://agentmods.dev/skills/junghan0611/entwurf/entwurf-dev)

Your own site

<a href="https://agentmods.dev/skills/junghan0611/entwurf/entwurf-dev"><img src="https://agentmods.dev/badge/skills/junghan0611/entwurf/entwurf-dev.svg" alt="Measured on agentmods" height="20"></a>

Per session 162 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 5,769 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.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00162	$0.05769
Opus 5	$0.00081	$0.02884
Sonnet 5	$0.00032	$0.01154
Haiku 4.5	$0.00016	$0.00577

Measured 3d ago against content hash be17bfbac118, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

entwurf-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 3d 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.

.claude/skills/entwurf-dev/SKILL.md · 335 lines

How it starts

The opening of the file, as written. The whole thing — 335 lines — stays where its author put it; the contents beside it link to each section on GitHub.

entwurf-dev — 개발 중인 제품을 GLG가 직접 겪는 손

Repository: ~/repos/gh/entwurf.

이 스킬은 GLG에게 명령문을 외우게 하지 않는다. GLG가 자연어로 의도를 말하면 에이전트가 현재 세션의 native entwurf_* tool을 호출하고, 각 receipt의 뜻과 다음 관측 지점을 짧게 설명한다. 형제는 disposable worker가 아니라 별도 runtime과 transcript를 가진 garden citizen이다.

호출 예

/skill:entwurf-dev status
/skill:entwurf-dev fresh pi openai-codex/gpt-5.6-luna 오늘 S0 상태를 한 문장으로 말해
/skill:entwurf-dev fresh claude-code claude-sonnet-5 README의 visible-first 계약을 읽어
/skill:entwurf-dev send <garden-id> 지금 상태를 답해줘
/skill:entwurf-dev resume <garden-id>
/skill:entwurf-dev tour pi
/skill:entwurf-dev boundary

자연어도 같은 뜻이다.

분신 하나 보이게 열어줘
새 Pi 형제를 열고 callback 뒤 인사까지 보내줘
살아 있는 Claude Code에게 이 메시지 전달해줘
아까 닫은 그 형제 다시 불러줘
지금 내가 직접 시험할 수 있는 경계를 보여줘

시작 전 runtime guard

호출된 tool schema가 worktree 문서보다 우선한다. 이 스킬은 현재 S1 계약에 맞는다.

entwurf_fresh_call backend는 정확히 pi | claude-code | copilot | omp이고 model은 required다. cwd는 선택 입력 하나: literal 절대경로(존재하는 디렉터리, #·trim·realpath 없음), 생략·""면 caller cwd에서 시작한다. cross-repo fresh 절 참조.
기본 정책은 Pi=openai-codex/gpt-5.6-luna, Claude Code=claude-sonnet-5다.
GLG가 “entwurf 소넷”이라고 하면 Pi + entwurf/claude-sonnet-5다.
Provider budget: sibling launch에 OpenRouter를 쓰지 않는다. 이는 GLG 개인의 embedding/image 전용 제한 rail이다. Claude Code 구독, Pi의 승인된 GPT/Codex·xAI 구독, 또는 direct endpoint로 이미 설정된 회사 API만 쓴다. model label은 billing rail 증거가 아니다. 요청된 model이 현재 OpenRouter로 resolve되면 launch·test turn·login check·probe script를 하지 말고 그 한 사실만 즉시 보고한다. GLG가 이미 승인한 rail의 형제를 요청하면 credential/login을 다시 묻거나 찾지 말고 fresh-call을 바로 한 번 호출한다.
entwurf_v2 intent는 정확히 fire-and-forget 하나다. 이 verb는 어떤 rail에서도 프로세스를 열지 않는다.
entwurf_peers는 사실 조회이며 생성·재개 명령이 아니다.
entwurf_resume_call은 입력이 target 하나다. record가 transcript·model·provider·cwd를 주므로 model override도, task도, prompt도 없다. 턴을 돌리지 않는다.

활성 tool schema가 owned-outcome을 노출하거나, entwurf_fresh_call이 없거나, fresh-call의 required model 필드가 없거나, entwurf_resume_call이 없다면 아무 launch/send도 하지 말고 “옛 extension이 로드됐다. Pi control session을 다시 열어야 한다”고 보고한다. owned-outcome은 절대 호출하지 않는다. worktree를 읽고 active runtime이 새 버전이라고 추정하지 않는다.

Read the full file on GitHub · 335 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.

3d ago Changed · +2 lines be17bfbac118
7d ago First seen · 333 lines · 162 tokens per session scan A 38d3256d0d08

Subscribe to this mod's changes

entwurf-dev is a skill published in the GitHub repository junghan0611/entwurf (27 stars, last pushed today), licensed MIT. It adds 162 tokens to every session and 5,769 once invoked, about $0.0008 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 2026-08-30.

Building agents Claude Code acp agent-client-protocol ai-agent codex gemini-cli

gemini-delegate

Use when [large codebase >200 lines], [web search current info], [adversarial review/critique].

davdittrich/gemini-delegate · 30 tokens

opentag

Deploy or operate OpenTag's supported paired setup when a user needs to bootstrap the self-hosted Docker Compose Control Plane and Slack Source App, configure or pair a local ACP Runner with a GitHub Project Target, start the Runner service, verify readiness, or diagnose a Slack mention that did not complete.

amplifthq/opentag · 64 tokens

configure-coddy

Change Coddy's own configuration when the user asks for it: edit settings, providers, models, logging, permissions, or find, install, update, and remove MCP servers and skills. Stages UCI-style commands and commits only after the user confirms saving. Load when the user explicitly asks to change a Coddy setting, or…

coddy-project/coddy-agent · 106 tokens

generate-rules

Scan the codebase and generate or refresh focused project rules under .coddy/rules/.

coddy-project/coddy-agent · 22 tokens

vulnerability-writeup

Turn vulnerability notes, disclosure reports, PoCs, source code, or Codex Security findings into self-contained, sceptically validated, natural-sounding vulnerability reports. Use for one vulnerability or a disclosure campaign; a Codex Security scan is optional.

bex-co/bex-security · 54 tokens

track-findings

Track validated Codex Security findings in Linear, Jira, GitHub issues, or draft GitHub security advisories. Use it for one finding or an explicitly selected batch of up to 25 findings tracked as Linear, Jira, or GitHub issues. Includes duplicate checks, exact previews, approval-gated writes, and readback. Do not use…

bex-co/bex-security · 79 tokens

entwurf: Skill for Claude Code