Token导航 LogoToken导航

Agent Skill · 待分类

beads-create(珠子创造)

beads-create 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

beads-create 是待分类类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为权限未标注;许可证信息还不明确,第三方扫描已通过,按安装量在待分类类的 11,209 个 Skill 里排第 4,973。

总安装 636周安装 26GitHub Stars 7下载 206核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:beads-create(珠子创造)
来源仓库:https://github.com/btraut/skills
仓库路径:skills/beads-create
安装命令:
npx skills add https://github.com/btraut/skills --skill beads-create
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/btraut/skills --skill beads-create

这个 Skill 是做什么的

这个 Skill 的分类还没最终确定,用法以来源仓库的说明为准。

beads-create 用于根据计划或规范创建 Beads 任务并执行评审环节,简化提交流程。自动检测当前安装的 tracker(bd 或 br),适配不同后端存储。适用于快速将 PRD 转化为可执行 issue,减少重复沟通。输出包含评审意见与待办项,供 implementers 直接认领处理。

适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/beads-create

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/btraut/skills --skill beads-create;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    权限没有明确标注,第一次使用时只给一个无关紧要的任务,观察它到底读了哪些文件和网络。

常见问题

6 条

beads-create 怎么安装?

在终端执行 npx skills add https://github.com/btraut/skills --skill beads-create。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

beads-create 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

beads-create 需要什么权限?

权限级别没有明确标注。第一次使用时建议只给无关紧要的任务,观察它实际读写了哪些内容。

beads-create 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

beads-create 用的人多吗?

公开数据是累计安装 636 次,近一周安装 26 次,下载 206 次,GitHub 7 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

beads-create 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理待分类类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

待分类类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
infsh-cli

infsh CLI · Codex、Claude

4,374,576382权限未标注未明确未收录
ai-video-generation

AI 视频生成 · Codex、Claude

4,313,376382调用外部服务未明确有提醒
lark-openapi-explorer

云雀 OpenAPI 浏览器 · Codex、Claude

2,064,1209,011需要联网未明确有提醒
lark-workflow-standup-report

Lark 工作流程站立报告 · Codex、Claude

2,041,6328,978需要联网未明确通过
use-my-browser

使用我的浏览器 · Codex、Claude

1,848,49140操作浏览器未明确未通过
beads-create

珠子创造 · 本页

6367权限未标注未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Beads Create

Use this skill to create Beads from a plan/spec and do the review pass so implementers can just pick up issues and ship. This skill must work with either classic Beads (bd) or beads_rust (br).

Workflow

  1. Resolve the tracker first

- If the user explicitly says bd, br, beads, or beads_rust, obey that. - Otherwise detect what is actually installed: - If only br is installed, use br. - If only bd is installed, use bd. - If both are installed, prefer the tracker that already owns the current workspace/history. Check the local .beads/ state and the existing workflow docs before guessing. - If both are installed and the workspace is still ambiguous, ask one targeted question instead of winging it. - Do not mix commands from both trackers in the same run.

  1. Load the right local guidance if needed

- Look for a loaded local skill on the user's machine that covers the chosen tracker and use that first. - For bd, prefer the local beads skill if it is available. - For br, prefer any loaded local br, beads_rust, or equivalent tracker skill if it is available. - If no local skill exists, rely on local CLI help and the installed tool itself: - bd --help, bd create --help, bd update --help - br --help, br create --help, br update --help - Do not send the user to public web docs from this skill. - For br, use --json on agent-facing commands whenever available.

  1. Decide which mode you are in

- If the user has a finalized plan/spec and wants Beads filed: do Create mode. - If the user already has Beads and wants them improved: do Review mode (skip straight to step 6).

  1. Confirm the plan is ready (Create mode)

- Ensure there is a finalized plan or spec outside Beads. - If the plan is still fuzzy, ask for revisions first and iterate up to 5 times before importing.

  1. Translate the plan into tracker structure (Create mode)

- Create epics that map to major milestones or deliverables. - Create issues for concrete, implementable tasks. - For bd: - When an epic has child milestones/tasks, create them as hierarchical children with dotted IDs like EPIC.1, EPIC.2, not as loose top-level issues. - Follow the bd create... --parent <epic-id> flow so bd auto-assigns EPIC.N IDs. - Verify the returned IDs. If bd does not assign dotted IDs, stop and ask the user how to proceed. - For br: - Use the epic/parent relationship, but do not require dotted child IDs. br supports parent-child structure without the old bd dotted-ID contract. - Create the epic first, then create or update child issues with --parent <epic-id>. - Add explicit dependencies where sequencing matters, and keep the graph acyclic with br dep cycles --json. - Add dependencies, ordering constraints, and opportunities for parallel work.

  1. Review and polish (Create mode + Review mode)

- Fix vague titles, missing acceptance criteria, or unclear scope. - Ensure each issue is something one agent can do in a single task. - Verify dependencies, sequencing, and opportunities for parallel work. - Add missing design notes / decision context. - Split oversized issues; merge duplicates; delete fluff. - For br, prefer structured updates over freeform ambiguity: - br update... --description - br update... --design - br update... --acceptance-criteria - br update... --notes - For bd, use the locally available bd fields/workflow for description, design, notes, and acceptance criteria on the installed version.

  1. Add implementation detail

- Include design notes, assumptions, and acceptance criteria on each issue. - Make tasks small and unambiguous so a fresh agent can pick up a single issue.

  1. Validate tracker state before leaving

- For br, run br dep cycles --json and make sure it is empty. - For br, remember sync is explicit. If you changed tracker state and need the JSONL export updated, run br sync --flush-only. - For bd, follow the local sync/export workflow supported by the installed version.

  1. Single-pass

- Do one focused pass in the current context. - If another pass is needed, say so explicitly and recommend a separate follow-up run.

  1. Restart the agent
  • Finish this task, then stop. Encourage a fresh agent for the next task to keep sessions small.

Output expectations

  • Provide a concise mapping of epics and issues you created.
  • State which tracker you used: bd or br.
  • Call out any dependencies or blockers that need user confirmation.
  • If asked to iterate beyond 5 plan revisions, say you do not think it can be improved further.

平台分布

Codex

32.74%
按下载量换算67

Claude

31.25%
按下载量换算64

Cursor

17.64%
按下载量换算36

Gemini CLI

10.34%
按下载量换算21

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

权限未标注

当前来源未能明确判断权限范围,默认进入异常复核队列。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills