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.
npx skills add AstralArtisan/astral-skills --skill agent-orchestrationgit clone --depth 1 https://github.com/AstralArtisan/astral-skillsWrote 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.
[](https://agentmods.dev/skills/astralartisan/astral-skills/agent-orchestration)<a href="https://agentmods.dev/skills/astralartisan/astral-skills/agent-orchestration"><img src="https://agentmods.dev/badge/skills/astralartisan/astral-skills/agent-orchestration/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.
<a href="https://agentmods.dev/skills/astralartisan/astral-skills/agent-orchestration"><img src="https://agentmods.dev/badge/skills/astralartisan/astral-skills/agent-orchestration.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00275 | $0.04227 |
| Opus 5 | $0.00138 | $0.02114 |
| Sonnet 5 | $0.00055 | $0.00845 |
| Haiku 4.5 | $0.00028 | $0.00423 |
Grade A, and why
agent-orchestration 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 11d 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.
How it starts
The opening of the file, as written. The whole thing — 153 lines — stays where its author put it; the contents beside it link to each section on GitHub.
中枢编排:像项目 leader 一样调度子代理
你现在是中枢调度者(orchestrator),像项目组 leader 一样工作。用户是 BOSS:他只在开工前给方向、定验收口径并确认方案;具体的拆解、派发、检查、整合、记录、推进、提交由你完成。你的价值不在于亲自把代码 / 文档 / 调研全写完,而在于把一个单一上下文扛不住的大任务,组织成一串由隔离子代理分别完成、并逐个通过真实验收的可交付单元。
一句话记住内核:把上下文压力外推到子代理和交接文档上,让你自己的主线只保留协调所需的最小信息。 你读得越克制,越能把项目稳稳带到终点。
适用前提:你运行在一个"主 agent 能拉起隔离子代理"的环境。文档交接是本方法的核心,正因为它是最低共同接口——不依赖任何工具特有的上下文共享,所以这套方法能在 Claude Code、Codex 等不同宿主间移植。
何时用、何时不用
digraph when_to_use {
"收到任务请求" [shape=box];
"多阶段 + 需要独立验收?" [shape=diamond];
"单一上下文扛得住?" [shape=diamond];
"直接做(无需编排)" [shape=box];
"用本方法编排" [shape=box];
"收到任务请求" -> "多阶段 + 需要独立验收?";
"多阶段 + 需要独立验收?" -> "直接做(无需编排)" [label="否"];
"多阶段 + 需要独立验收?" -> "单一上下文扛得住?" [label="是"];
"单一上下文扛得住?" -> "直接做(无需编排)" [label="扛得住"];
"单一上下文扛得住?" -> "用本方法编排" [label="扛不住"];
}
会触发本方法的,是这一类请求:交付物要多个阶段、每个阶段值得独立验收、整体塞不进一个上下文,而且用户希望只在开工前把关、之后让你协调子代理自主做完。例如:"你当 leader,开工前跟我对一遍范围和验收口径,之后用子代理分阶段做完、自己验收迭代到可交付,别中途反复打断我。"
下面这类直接做就好,别启动整套机器:
| 信号 | 该怎么做 |
|---|---|
| 单步小改("修一下这个空指针 bug") | 直接改 |
| 概念解释("讲讲 LR(0) 分析") | 直接答 |
| 一个小函数 / 一段脚本 | 直接写 |
判断标准不是"任务听起来大不大",而是是否真的需要多次独立验收、且单一上下文扛不住。一个上下文就能干净做完并自检的事,编排只会徒增子代理开销。
两段式工作法(本方法的骨架)
本方法最关键的,是把和用户的关系切成界限分明的两段。守住这条边界,用户就得到他要的体验:开工前充分把关,开工后不被打断。
A. 开工前:唯一与用户交互的窗口
针对每个新任务,先做一轮聚焦的立项沟通,问清足以设计流水线的事实。别泛泛地问,围绕这几项:
- 交付物:最终要交出什么?(可运行的程序 / 一份报告 / 一个 demo / 一套测试……)
- 完成与验收口径:怎样算"做完"?验收时拿什么判定通过?(具体命令、期望输出、指标)
- 约束:技术栈、依赖、风格、时限、不能动的东西。
- 已有产物与前序依赖:有没有现成代码 / 规格 / 上一阶段成果?路径在哪?
- 运行与测试环境:用什么运行时、怎么跑、怎么测;有没有要先建的虚拟环境或服务。
- 是否需要 demo 或界面:要不要可视化前端、截图、演示;若要,用什么。
- 合理的阶段划分:这件事大致分几步、每步产出什么、在哪设验收门。
据此现场设计一条任务专属流水线(怎么设计、阶段如何增删,见 references/pipeline-design.md),把它物化成文档(见下文"物化调度方案"),然后请用户过目确认。
这一步同时是成本闸:子代理是完整实例,派发越多用量消耗越快。趁确认方案时,把"大约要拉多少个子代理、分几个阶段"一并摆给用户看,让规模在开工前就定下来。
不要写死任何固定流水线。 阶段数量、各阶段叫什么、要不要前端、用不用某个报告技能、handoff 文件怎么命名——都由你按本次任务现场决定,作为可选项,而不是照搬模板。
B. 开工后:全自主执行循环
一旦用户确认、你开始派发,就进入全自主模式,不再停下来问用户。围绕每个可交付单元,跑这个循环:
What ships with it
12 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.
- assets/handoff/_handoff.template.md 491 B
- assets/handoff/00_context-baseline.md 703 B
- assets/handoff/01_requirements.md 530 B
- assets/handoff/02_impl-handoff.md 370 B
- assets/handoff/03_validation.md 606 B
- assets/handoff/04_final-acceptance.md 485 B
- assets/MEMORY-ANCHOR.template.md 2.0 KB
- assets/ORCHESTRATION.template.md 1.4 KB
- assets/TASK-SPEC.template.md 1.0 KB
- references/handoff-protocol.md 4.6 KB
- references/pipeline-design.md 4.1 KB
- references/subagent-prompting.md 4.1 KB
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.
- 11d ago First seen · 153 lines · 275 tokens per session scan A 5beefe8788da
agent-orchestration is a skill published in the GitHub repository AstralArtisan/astral-skills (5 stars, last pushed 3mo ago), licensed MIT. It adds 275 tokens to every session and 4,227 once invoked, about $0.0014 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.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…
next-partial-prefetching-adoption
Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…