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 xz1220/open-dynamic-workflows --skill zh-cngit clone --depth 1 https://github.com/xz1220/open-dynamic-workflowsWrote 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/xz1220/open-dynamic-workflows/zh-cn)<a href="https://agentmods.dev/skills/xz1220/open-dynamic-workflows/zh-cn"><img src="https://agentmods.dev/badge/skills/xz1220/open-dynamic-workflows/zh-cn/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/xz1220/open-dynamic-workflows/zh-cn"><img src="https://agentmods.dev/badge/skills/xz1220/open-dynamic-workflows/zh-cn.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00154 | $0.18283 |
| Opus 5.5 | $0.00062 | $0.07313 |
| Sonnet 5.5 | $0.00031 | $0.03657 |
| Haiku 4.5 | $0.00015 | $0.01828 |
Grade A, and why
open-dynamic-workflows 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 6d 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 — 730 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Open Dynamic Workflows
一个 dynamic workflow 就是一段简短的 JavaScript 脚本：编排计划写成普通代码，由
odw 在独立的后台进程里执行，把每个子任务派发给一个真实的 coding-agent CLI 进程。
中间产物不进入你的上下文，回到你手里的只有脚本 return 的最终值。
使用流程固定三步：写脚本 → odw run → 检视结果再行动。任务一次调用就能完成时
不要用——直接做。
写 workflow 脚本
- 文件最顶部放
export const meta = {…}，必须是纯字面量（不含变量、函数调用、 模板插值）。meta.name与meta.description必填；whenToUse、phases、model可选。phases的条目是{ title, detail? }（或者只写标题）；条目里的adapter和model是该阶段 agent 的默认值（ODW 扩展——见按阶段路由）。 - 脚本体运行在 async 上下文里：直接用顶层
await；顶层return的值就是整个 workflow 的结果。 - 原语全部是注入的全局——不要 import。除
export const meta外，文件里出现任何 其他顶层import/export都会被加载器拒绝。 - 循环、
if、去重等普通控制流直接写在脚本里。原语只负责派发并等待，拿结果做 什么由脚本决定。 - 要同时能在 Claude Code 的 Workflow 工具上运行，就只用方言本身：不用 ODW 扩展，也不依赖
一致性测试套件标为
unspecified的行为。通过一致性测试的脚本两边都能跑——odw conformance运行该套件（ODW 仓库的conformance/；见原语参考）。
// fan-out-reduce.js
export const meta = {
name: 'fan-out-reduce',
description: 'Draft in parallel, then synthesize.',
phases: [{ title: 'Draft' }, { title: 'Synthesize' }],
}
const question = (args && args.question) || 'Design a cache.'
phase('Draft')
const drafts = await parallel(
[1, 2, 3].map((i) => () => agent(`Draft #${i}: ${question}`, { phase: 'Draft' })),
)
phase('Synthesize')
return await agent(
'Synthesize the best answer from:\n' + drafts.filter(Boolean).join('\n---\n'),
{ phase: 'Synthesize' },
)
用 --args 传入的输入会注入为全局 args（解析后的 JSON，或一段原始字符串——
看起来像 JSON 但解析失败的输入会被直接拒绝，不会悄悄按字符串传入）。
原语速查
| 原语 | 作用 |
|---|---|
agent(prompt, opts?) |
让一个 coding agent 跑一个子任务；返回它的回复文本，设了 opts.schema 则返回校验过的对象。唯一真正干活的动词。 |
parallel(thunks) |
并发执行一组零参 thunk，并等全部完成（屏障）。顺序保留；失败的槽位是 null。parallel(thunks, { notes: true }) 让其 agent 使用本次运行的笔记板（ODW 扩展；见下文）。 |
pipeline(items, ...stages) |
让每个条目独立地流过各 stage（无屏障）。每个 stage 收到 (prev, item, index)。 |
phase(title, opts?) / log(msg) |
给后续工作打上进度标签 / 发一行进度消息。opts（{ adapter, model }，ODW 扩展）为该阶段的 agent 指定 CLI 和模型。 |
args |
workflow 的输入（注入）。 |
budget |
{ total, spent(), remaining() }——按 token 目标扩缩深度。 |
workflow(ref, args?) |
内联调用另一个 workflow（仅一层）。ref 是受管目录中的名字或 { scriptPath }；子 workflow 共享本次运行的并发上限、agent 计数和预算。 |
validate(source) |
解析并编译检查源码,不运行工作流正文;meta 按纯字面量解析,绝不求值。返回 { ok, meta?, errors, warnings }。ODW 扩展——不属于 Claude Code 方言。 |
exec(command, opts?) |
不启动模型直接运行一条命令（测试、构建、linter）；返回 { stdout, stderr, exitCode, signal, timedOut, truncated, durationMs }。字符串经 shell 执行，数组按 argv 执行。ODW 扩展——不属于 Claude Code 方言。 |
checkpoint(title, opts?) |
暂停运行，直到有人批准（可以附上修改后的 payload）或拒绝；返回 { approved, note, by, payload, edited }。ODW 扩展——不属于 Claude Code 方言。 |
What ships with it
2 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.
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.
- 6d ago Changed · +20 lines 78b6db3167bc
- 7d ago Changed · +332 lines dc5990cfad12
- 8d ago Changed · +242 lines 990dd9b34238
- 1mo ago Changed · +1 lines a8374e9deb80
- 1mo ago First seen · 135 lines · 154 tokens per session scan A 36fd354f3c50
open-dynamic-workflows is a skill published in the GitHub repository xz1220/open-dynamic-workflows (112 stars, last pushed 7d ago), licensed MIT. It adds 154 tokens to every session and 18,283 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 2026-08-30.
Other skills, from other repositories
gpui-mockup
Explore visual directions for a herdr-gpui UI element as real GPUI code. Write 3–4 variants in one Rust file, show them side by side in a native mockup window in the user's own theme and fonts, capture the window to check your work, and iterate on the picks and notes the user sends back. Use instead of the HTML mockup…
release
Publish a herdr-gpui release. Use when the user asks to cut, push, ship, or publish a release, or to approve, verify, or GPG-sign one that is already running. Covers just release, the protected environments the run stops at, and the local verification and signing steps that are not part of CI.
herdr-gpui-browser
Open a browser tab in Herdr GPUI to show the user a web page or an HTML file you wrote, next to the terminal you run in, and get back the notes the user pins on it. Use whenever the user asks to open, start, or show a browser tab, a page, a preview, or a mockup while you run inside Herdr (HERDRENV=1 is set in your…
skill
Skill "skill" from wavyrai/tmux-ide, covering tmux-ide — claude code skill, when to use, the agent contract, coordinating with other agents and fleet control from the cli.
taskflow
A workflow coordinator for splitting large coding tasks into multiple phases handled by separate agents.
supermux-core
Use when building an app, script, bot or service in TypeScript that drives Claude Code, Codex, Cursor, OpenCode, Grok Build or Antigravity through the @supermux/core library.