Token导航 LogoToken导航

Agent Skill · 研究检索

workflow-start(工作流程启动)

workflow-start 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。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。

workflow-start 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 66,030。

总安装 269周安装 11GitHub Stars 6下载 86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:workflow-start(工作流程启动)
来源仓库:https://github.com/duc01226/easyplatform
仓库路径:skills/workflow-start
安装命令:
npx skills add https://github.com/duc01226/easyplatform --skill workflow-start
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/duc01226/easyplatform --skill workflow-start

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。支持基于关键词、任务场景或来源线索进行信息聚合与过滤,提升研究效率。通过 npx 命令从 GitHub 仓库安装,具体用法需结合原始 README 进一步确认。

安装前建议核实权限范围、维护状态及是否涉及联网、命令执行或文件操作。workflow-start 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/workflow-start

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/duc01226/easyplatform --skill workflow-start;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

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

workflow-start 怎么安装?

在终端执行 npx skills add https://github.com/duc01226/easyplatform --skill workflow-start。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

workflow-start 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。

workflow-start 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

workflow-start 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。

workflow-start 用的人多吗?

公开数据是累计安装 269 次,近一周安装 11 次,下载 86 次,GitHub 6 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。

workflow-start 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。
  • 标注为只读权限,试用阶段不容易误改本地文件。

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。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

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。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
workflow-start

工作流程启动 · 本页

2696只读未明确通过

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

[MANDATORY] You MUST ATTENTION use TaskCreate to break ALL work into small tasks BEFORE starting. NEVER skip task creation.

Quick Summary

Goal: Activate a workflow by ID, creating tracking tasks and announcing the workflow sequence.

Workflow:

  1. Match -- Validate workflow ID against catalog
  2. Create -- Set up TaskCreate items for all workflow steps
  3. Announce -- Tell user the detected intent and workflow sequence

Key Rules:

  • AI MUST ATTENTION always detect nearest workflow and ask user via AskUserQuestion to confirm activation
  • Present "Activate [Workflow] (Recommended)" vs "Execute directly without workflow"
  • Create workflow-level tasks BEFORE any implementation tasks

Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).

Workflow Start

Activate a workflow from the injected catalog and initialize step tracking via TaskCreate.


When to Use

  • Starting a detected workflow from the injected catalog
  • Initializing workflow state for step-by-step execution
  • Switching from one active workflow to another

NOT for: Manual step execution (follow TaskCreate items), workflow design (use planning), or workflow catalog management.


Quick Reference

Workflow

  1. Read the workflow catalog injected by the hook
  2. Validate the workflow ID against available entries
  3. Create workflow state via the workflow-step-tracker hook
  4. Create TaskCreate items for ALL sequence steps immediately
  5. Begin first step, marking it in_progress

Related

  • Command: /workflow-start <workflowId>
  • Hook: workflow-step-tracker.cjs
  • Hook: workflow-router.cjs

Activation Rules

  • Call ONLY after reading the workflow catalog injected by the hook
  • Use the exact workflow ID shown in the catalog (e.g., feature, bugfix, investigation)
  • ALWAYS ask the user via AskUserQuestion to confirm activation before proceeding
  • If called while another workflow is active, it will auto-switch (end current, start new)

After Activation — Task Creation Protocol (ZERO TOLERANCE)

Your FIRST action after activation MUST ATTENTION be creating EXACTLY one TaskCreate for EACH entry in the workflow's sequence array from .claude/workflows.json.

How to determine the task list:

  1. Read .claude/workflows.json
  2. Find the activated workflow by ID
  3. Read its sequence array — this is the SOLE source of truth
  4. Create one TaskCreate per array entry, IN ORDER

Task format:

TaskCreate: subject="[Workflow] /{step-name} — {brief description}", description="Workflow step N/{total}. {conditional note if applicable}", activeForm="Executing /{step-name}"

Rules (NON-NEGOTIABLE):

  • 1:1 mapping — Each sequence entry = exactly one task. No consolidation, no summarization, no custom task names.
  • Conditional steps still get tasks — Steps that may be skipped at runtime (e.g., /plan, /cook after a PASS review) STILL get created as tasks. Add to description: "Conditional — skip if reviews pass".
  • Recursive self-calls get tasks — If the sequence references itself (e.g., workflow-review-changes within review-changes), create a task: [Workflow] /workflow-review-changes — Recursive re-review (conditional).
  • Never invent tasks — Do NOT add tasks not in the sequence array (no "Assess results", no "Final review", no "Check output"). Only create what the sequence defines.
  • Count verification — After creating all tasks, verify: task count == len(sequence). If mismatch, fix immediately before proceeding.

Create ALL tasks first, then mark the first task in_progress via TaskUpdate.

Step Execution Protocol (applies to ALL workflows)

Per step: TaskUpdate in_progress → invoke Skill tool → complete skill workflow → TaskUpdate completed.

Marking a task completed without invoking its Skill tool is a workflow violation. Validation gates (/plan-validate, /plan-review, /why-review) require AskUserQuestion — never auto-approve or batch-complete them.

To skip a conditional step: TaskUpdate in_progress → add comment "Skipped — {reason}" → TaskUpdate completed. Do NOT delete conditional tasks.


See Also

  • planning skill - Creating implementation plans
  • feature-implementation skill - Implementing features after planning
  • debug-investigate skill - Bug fix workflows

IMPORTANT Task Planning Notes (MUST ATTENTION FOLLOW)

  • Always plan and break work into many small todo tasks
  • Always add a final review todo task to verify work quality and identify fixes/enhancements

Closing Reminders

  • IMPORTANT MUST ATTENTION break work into small todo tasks using TaskCreate BEFORE starting
  • IMPORTANT MUST ATTENTION search codebase for 3+ similar patterns before creating new code
  • IMPORTANT MUST ATTENTION cite file:line evidence for every claim (confidence >80% to act)
  • IMPORTANT MUST ATTENTION add a final review todo task to verify work quality

平台分布

Codex

33.88%
按下载量换算29

Claude

29.99%
按下载量换算26

Cursor

19.38%
按下载量换算17

Gemini CLI

8.5%
按下载量换算7

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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