ai-bouncer: Skill for Claude Code

.agents/skills/dev-bounce/SKILL.md

dev-bounce is a skill for Claude Code, Codex from kangraemin/ai-bouncer. It costs 138 tokens per session (1,511 once invoked), scanned A, original, MIT.

A structured workflow for code changes, feature work, bug fixes, refactoring, and file edits.

In plain words
What is it for?
Use it whenever you modify code, implement functionality, fix bugs, refactor, or change project files.
Why use it?
It gives development tasks a consistent sequence of phases and steps to follow.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: names the AskUserQuestion tool; installed under .agents/ (shared by several agents).

This is kangraemin/ai-bouncer's own configuration. It tells Claude Code and Codex how to work on ai-bouncer itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything ai-bouncer configures →

Reuse

Borrowing it

Nothing to install: this file belongs to kangraemin/ai-bouncer. 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/kangraemin/ai-bouncer/main/.agents/skills/dev-bounce/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/kangraemin/ai-bouncer

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 dev-bounce

README.md
[![agentmods](https://agentmods.dev/badge/skills/kangraemin/ai-bouncer/dev-bounce/github.svg)](https://agentmods.dev/skills/kangraemin/ai-bouncer/dev-bounce)
Your own site
<a href="https://agentmods.dev/skills/kangraemin/ai-bouncer/dev-bounce"><img src="https://agentmods.dev/badge/skills/kangraemin/ai-bouncer/dev-bounce/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 dev-bounce

Your own site · 80×15
<a href="https://agentmods.dev/skills/kangraemin/ai-bouncer/dev-bounce"><img src="https://agentmods.dev/badge/skills/kangraemin/ai-bouncer/dev-bounce.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 138 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,511 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00138 $0.01511
Opus 5.5 $0.00055 $0.00604
Sonnet 5.5 $0.00028 $0.00302
Haiku 4.5 $0.00014 $0.00151

Measured 4d ago against content hash 53e4d7e3ac26, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

dev-bounce 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 4d 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.

.agents/skills/dev-bounce/SKILL.md · 104 lines

How it starts

The opening of the file, as written. The whole thing — 104 lines — stays where its author put it; the contents beside it link to each section on GitHub.

dev-bounce

작업을 레포 설정(workflow.yaml)에 적힌 모드의 단계 순서대로 진행한다. 단계마다 할 일(prompt)은 엔진이 주입하고, 통과 조건(run)은 엔진이 실행해 판정한다. 마지막 단계까지 통과하기 전에는 멈출 수 없다.

이 문서에는 시작 절차만 있다. 시작한 뒤에는 주입되는 지시를 따른다.

Step 1 — 상태 확인

bouncer scan
STATE    MINE <경로> <모드> <단계>         이 세션이 이어서 할 작업
STATE    NONE                             없음
PICK     <방식>                           모드를 고르는 방식 (ask / auto)
WORKFLOW <모드> <설명>                     모드 선택지
CHOICE   <모드> <단계> <id> <방식> <이름>   시작할 때 골라야 할 항목 (방식 ask / auto)
ERROR    <메시지>                          설정 문제 — 사용자에게 그대로 알린다
  • MINE이면 이미 진행 중이다. 출력 맨 끝에 붙은 현재 단계 지시를 따라 이어서 한다.
  • NONE이면 Step 2로 간다.
  • bouncer: command not found가 나오면 이 컴퓨터에 설치되지 않은 것이다. 사용자에게 알리고 평소대로 작업한다.

Step 2 — 개발 작업인지 판별

  • 질문·설명·코드 읽기면 스킬을 쓰지 않고 답한다.
  • 무엇을 고칠지 불명확하면 AskUserQuestion으로 구체화한다.
  • 개발 요청이면 Step 3으로 간다.

Step 3 — 모드와 항목 정하기

  1. 모드를 정한다. 방식은 PICK 줄에 있다.
    • ask면 WORKFLOW 줄로 AskUserQuestion을 만들어 사용자에게 묻는다. 모드는 사용자가 고른다.
    • auto면 요청 내용을 보고 직접 고른다. 무엇을 왜 골랐는지 한 줄로 알린다.
  2. 고른 모드의 CHOICE 줄을 단계별로 묶어 처리한다.
    • 방식이 ask면 반드시 사용자에게 묻는다(AskUserQuestion, multiSelect). 선택지가 4개를 넘으면 질문을 나눈다.
    • 방식이 auto면 요청 내용을 보고 직접 고른다. 무엇을 왜 골랐는지 한 줄로 알린다.
    • CHOICE가 없으면 건너뛴다.

AskUserQuestion을 쓸 수 없는 세션(서브에이전트, claude -p)이면 추측해서 진행하지 않는다. 선택지를 텍스트로 제시하고 답을 기다린다.

Step 4 — 시작

bouncer start <모드> "<slug>" --pick <단계>=<id>,<id> [--pick <단계>=…]
  • slug는 영문 소문자와 하이픈으로 쓴다(예: payment-bug).
  • 고르는 단계는 전부 --pick으로 적어야 한다. 하나도 고르지 않은 단계는 --pick <단계>=로 적고, 그 단계는 건너뛴다.
  • 출력에 첫 단계의 지시가 나온다. 그대로 따른다.

Step 5 — 이후

  • 끝내려 하면 엔진이 그 단계의 run을 돌린다.
    • 실패하면 출력을 보여 주며 계속하게 한다. 원인을 고친다.
    • 통과하면 다음 단계로 넘어가며 다음 지시를 준다.
  • 마지막 단계까지 통과하면 끝이다.
  • 단계는 스스로 넘길 수 없다. 상태·설정 파일 편집은 막힌다.
  • 같은 실패가 반복되면 같은 수정을 되풀이하지 말고 다른 원인을 찾는다.
    • 연속으로 여러 번 막히면 엔진이 사용자에게 판단을 넘기라고 안내한다.
  • 작업을 접을 때는 bouncer cancel을 쓴다.

명령

명령 용도
bouncer scan 상태(이어서 할 작업이면 현재 단계 지시)·모드·고를 항목
bouncer start <모드> <slug> --pick … 작업 시작
bouncer cancel 작업 취소
bouncer check workflow.yaml 검사
bouncer record <이름> <명령> 검사를 직접 돌리고 결과를 남긴다 (prompt가 시키면 이걸로 돌린다)
bouncer passed <이름> run용: 그 결과가 최신이고 통과했는지 (아니면 exit 2)
bouncer init [--defaults|--decline] 레포에 켜기(빈 설정) / 기본값 써 넣기 / 켜지 않기
bouncer setup 레포 설정을 사용자와 대화로 정한다 — 출력된 안내대로 묻고 yaml 을 쓴다

Read the full file on GitHub · 104 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. 4d ago Changed 53e4d7e3ac26
  2. 5d ago Changed · -111 lines · -22 tokens per session 65b355cbe2bf
  3. 7d ago Changed 3acd917e5e43
  4. 24d ago Changed · -5 lines 8a51f9bf2d1d
  5. 27d ago Changed · -763 lines · +53 tokens per session f2cb97a19cb2
  6. 1mo ago First seen · 983 lines · 107 tokens per session scan A ee63a8ff1035

Subscribe to this mod's changes

dev-bounce is a skill published in the GitHub repository kangraemin/ai-bouncer (3 stars, last pushed 4d ago), licensed MIT. It adds 138 tokens to every session and 1,511 once invoked, about $0.0006 per session on Opus 5.5. 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

issue-triage

3-phase issue backlog management with audit, deep analysis, and validated triage actions. Use when triaging GitHub issues, sorting bug reports, cleaning up stale tickets, or detecting duplicate issues. Args: 'all' to analyze all, issue numbers to focus (e.g. '42 57'), 'en'/'fr' for language, no arg = audit only.

FlorianBruniaux/claude-code-plugins · 81 tokens

review-pr

Perform a comprehensive code review of a pull request.

FlorianBruniaux/claude-code-plugins · 12 tokens

scaffold

Interactive coach that asks 4-5 questions to determine whether you need an agent, command, skill, hook, or rule, then generates a ready-to-use template. Usage: /scaffold (no arguments needed, starts the coaching session).

FlorianBruniaux/claude-code-plugins · 52 tokens

security-audit

Comprehensive security audit with scored posture assessment.

FlorianBruniaux/claude-code-plugins · 12 tokens

guide-recap

Transform CHANGELOG entries into social content (LinkedIn, Twitter/X, Newsletter, Slack) in FR + EN. Use after releases or weekly to generate release notes, announcements, social media posts, or recap summaries from guide updates.

FlorianBruniaux/claude-code-plugins · 50 tokens

release-notes-generator

Generate release notes in 3 formats (CHANGELOG.md, PR body, Slack announcement) from git commits. Automatically categorizes changes and converts technical language to user-friendly messaging. Use for releases, changelogs, version notes, what's new summaries, or ship announcements.

FlorianBruniaux/claude-code-plugins · 59 tokens