Codex Dispatch
Mandatory dispatch preflight before any subagent delegation. Use before the first child-work call in every task, and when starting Codex work that needs an explicit Direct vs Single/Parallel/Sequential choice: code edits, debugging, reviews, repo exploration, multi-file changes, ambiguous scope, high-risk judgment, or requests mentioning agents, delegation, parallel work, model choice, cost, token use, or over-delegation.
Install
npx skills add yinguangyao/codex-dispatch-skill --skill codex-dispatchgit clone --depth 1 https://github.com/yinguangyao/codex-dispatch-skill.git /tmp/codex-dispatch-skill
mkdir -p ~/.claude/skills
cp -r /tmp/codex-dispatch-skill/. ~/.claude/skills/codex-dispatchPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# Codex Dispatch ## Mandatory preflight Before the first call that creates or assigns substantive work to any child agent in a user task, the root MUST load this skill and emit the visible dispatch decision below. This is a hard gate: without the decision, do not call `spawn_agent`, `followup_task`, or send a scope-expanding task to an existing child. One decision covers later status checks, waits, messages within the approved scope, interrupts, and result collection for that task. A materially new user request or a new child scope requires a new preflight before more delegation. If this skill is unavailable, stay Direct or tell the user delegation is blocked. If a child was started before the preflight, stop further delegation, disclose the miss, load this skill, emit the decision, and audit every existing child against the chosen mode, scope, and three-child maximum. Keep conforming children; interrupt or rescope non-conforming children. Do not restart conforming work merely to simulate compliance. ## Core rule Make one visible dispatch decision before substantial work and always before child delegation. Loading this skill never means “spawn”; it means decide. Root owns requirements, shared context, file-conflict control, final synthesis, and the user-facing answer. Delegate only when a child result is independently useful enough to justify coordination overhead.
Details
- Skill name
codex-dispatch- Repository
- yinguangyao/codex-dispatch-skill
- Repo updated
- 19 July 2026
FAQ
What does the Codex Dispatch skill do?
Mandatory dispatch preflight before any subagent delegation. Use before the first child-work call in every task, and when starting Codex work that needs an explicit Direct vs Single/Parallel/Sequential choice: code edits, debugging, reviews, repo exploration, multi-file changes, ambiguous scope, high-risk judgment, or requests mentioning agents, delegation, parallel work, model choice, cost, token use, or over-delegation.
How do I install the Codex Dispatch skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add yinguangyao/codex-dispatch-skill --skill codex-dispatch. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Codex Dispatch popular?
It lives in yinguangyao/codex-dispatch-skill, which has 5 GitHub stars and 0 forks.
Related skills
Ddd Architecture Coach
5jed1978/ddd-architecture-coach
DDD Architecture Coach — a decision-making coach spanning DDD (Strategic + Tactical Patterns), AI/LLM engineering (intervention design, risk assessment, fallbacks), software engineering discipline (Clean/Hexagonal Architecture, testing, CI/CD, SBE), and cloud architecture (containers, serverless, observability, cost). Runs a four-phase architecture planning workflow (Phase 1 Domain Discovery / Phase 2 Architecture Design / Phase 3 Implementation Spec / Phase 4 Review & Iterate) and produces Context Maps, Aggregate designs, AI-ADRs, Key Examples (Gherkin), and decision records — NOT code. Suppo
Camelot
5Barinfo/Camelot.skill
No-mercy code review in the voice of Arthurian judgment. Code either proves itself worthy — like drawing the Sword from the Stone — or is rejected without sentiment. Evidence-bound findings, banishment-level severity grading (Exile/Dishonor/Counsel/Valour), and a final verdict at Camlann with actionable reforging orders. Use when the user asks for a harsh, expert code/PR review, or an evaluation of in-progress agent work. Stateless single-pass judge: it reviews whatever round of code it is handed; recursive round-robin is delegated to Ouroboros.
Libby Book Monitor
5alexpolonsky/agent-skill-libby-book-monitor
Track book availability on Libby/OverDrive libraries. Search library catalogues, manage a watchlist, and get notified when books are added. Use for "libby", "check libby", "libby watchlist", "is book on libby", "book available", "overdrive", "library catalogue", "ספרייה", "ספרים".
Book Distiller
5AINoSleep/book-distiller-skill
炼金书:输入书名或电子书,自动深度调研→方法论提炼→生成该书的可运行框架Skill。 两种入口:(1)明确书名→直接蒸馏 (2)模糊学习需求→诊断荐书→再蒸馏。 触发词:「蒸馏这本书」「炼金XX」「把XX这本书做成skill」「XX的方法论skill」「更新XX的书skill」。 模糊需求也触发:「我想系统学谈判」「有没有一本书能帮我提升XX」。 English triggers: "distill [book]", "turn [book] into a skill", "book to framework skill". 不在用户只要书摘、读书笔记、买书推荐或查某个概念时触发——只在需要生成「可运行的方法论Skill」时激活。
Mailbot Programmable Inbox
5yopiesuryadi/mailbot-programmable-inbox
Integrate mailbot API for AI agent email automation. Create programmable inboxes, send and receive emails, track delivery events, and handle email-based workflows via API, SDK, or MCP Server. Use when building AI agents that need email identity, when setting up agent-to-human communication, when replacing Gmail or Outlook with programmatic email, or when any task involves sending, receiving, or processing emails for AI agents. Covers Node.js SDK, Python SDK, and MCP Server for Claude integration. Always use this skill when the user mentions mailbot, programmable inboxes, agent email, or email
Bambu 3mf
5woolimi/bambu-3mf
코드로 만든 STL 부품을 Bambu Lab P2S 로 바로 뽑을 수 있는 Bambu Studio 프로젝트(.3mf)로 만든다. 부품마다 역할·하중·맞물림·환경을 정리한 부품 카드에서 재질(PLA·PETG·TPU·ABS·CF), plate, process preset, 부품별 설정(벽·infill·support·brim), 출력 방향, plate 배치를 Bambu Studio 번들 profile·Bambu Lab Wiki 근거로 정하고, Studio CLI 로 평탄화한 system profile 을 넣어 export → mesh·설정 검사 → plate 별 slice(시간·무게)까지 한 번에 돌린다. "3MF 만들어줘", "출력 파일 만들어줘", "Bambu Studio 프로젝트로", "바로 뽑을 수 있게", "슬라이스 설정 넣어줘", "필라멘트 뭐로 뽑아", "PETG 로 바꿔", "벽 늘려줘", "출력 시간 얼마나 걸려", "이 3MF 검토해줘", "설정 맞는지 봐줘", make_3mf·build_bambu_project 스크립트를 새로 만들거나 고칠 때, 이미 만든 3MF 를 검토할 때(scripts/review_3mf.py) 쓴다. 형상 정리·설계 뷰
Browse the most popular Claude skills, all software development skills, or write your own with the skills generator.