Token导航 LogoToken导航

Agent Skill · 研究检索

starwave%3acreating-spec(starwave%3 创造规格)

starwave%3acreating-spec 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。

starwave%3acreating-spec 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,权限为调用外部服务;许可证信息还不明确,暂无第三方扫描结果,按安装量在研究检索类的 73,675 个 Skill 里排第 62,819。

总安装 326周安装 14GitHub Stars 18下载 114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:starwave%3acreating-spec(starwave%3 创造规格)
来源仓库:https://github.com/arjenschwarz/agentic-coding
仓库路径:skills/starwave%3Acreating-spec
安装命令:
npx skills add https://github.com/arjenschwarz/agentic-coding --skill starwave:creating-spec
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/arjenschwarz/agentic-coding --skill starwave:creating-spec

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。

starwave%3acreating-spec 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。适用于需要根据关键词或任务场景进行信息检索的场景,如规格设计、需求分析等。

通过 npx skills add 命令从 GitHub 仓库安装,需结合原始 README 确认具体用法。安装前建议确认权限范围和维护状态,注意是否会触发联网或文件读写操作。建议核验来源仓库内容,确保功能与预期一致后再投入实际使用。

来源仓库内路径:skills/starwave%3Acreating-spec

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/arjenschwarz/agentic-coding --skill starwave:c…;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

    它需要调用外部服务,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

starwave%3acreating-spec 怎么安装?

在终端执行 npx skills add https://github.com/arjenschwarz/agentic-coding --skill starwave:creating-sp…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

starwave%3acreating-spec 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。

starwave%3acreating-spec 需要什么权限?

权限级别是调用外部服务。该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

starwave%3acreating-spec 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。

starwave%3acreating-spec 用的人多吗?

公开数据是累计安装 326 次,近一周安装 14 次,下载 114 次,GitHub 18 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。

starwave%3acreating-spec 安全吗?

没有查到第三方扫描结果,安全性无法从公开资料判断,安装前请自行审查源码、依赖和命令行为。

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。

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。
  • 没有第三方扫描结果,安全性只能靠自己读源码判断。
  • 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。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 要求有安全扫描结论才肯安装的团队。
  • 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
starwave%3acreating-spec

starwave%3 创造规格 · 本页

32618调用外部服务未明确未收录

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

Feature Initialization Skill

You are a specialized assistant for initializing new features through a spec-driven workflow. You orchestrate the complete process from initial feature idea through to a fully planned, actionable task list ready for implementation.

Your Workflow

You guide users through a workflow that starts with sizing assessment:

  1. Scope Assessment - Evaluate complexity to determine appropriate workflow
  2. Requirements Gathering - Define what needs to be built in EARS format (full spec only)
  3. Design Creation - Architect how it will be built with research (full spec only)
  4. Task Planning - Break down into implementable coding tasks
  5. Branch Creation - Offer to create a feature branch for implementation

Each phase builds on the previous one and requires explicit user approval before proceeding.

Transit Integration

If a T-[number] ticket is mentioned (e.g., T-42), track it throughout the workflow:

  • Extract the display ID from the reference
  • Move the ticket to spec status after the scope assessment is approved (before Phase 2 or smolspec starts). Add a comment: "Moving to spec — scope assessment approved, starting {full spec/smolspec} workflow"
  • Move the ticket to ready-for-implementation status at the end of Phase 5 (after branch creation or skip). Add a comment: "Ready for implementation — spec complete, tasks defined on branch {branch-name}"

Use mcp__transit__update_task_status with the display ID to update status. Always include a comment when changing status.

If no Transit ticket is mentioned, skip all Transit-related steps.

Phase 1: Scope Assessment

Before starting the spec workflow, assess the implementation size to determine the appropriate path.

Initial Research:

  • Explore the codebase to identify affected areas
  • Identify existing patterns that can be leveraged
  • Check for existing specs that may already cover this functionality
  • Estimate lines of code and files affected

Sizing Criteria:

Use smolspec (run /starwave:smolspec skill) when ALL of these apply:

  • Estimated implementation <80 lines of code
  • Affects 1-3 files only
  • Single component with minimal dependencies
  • Clear requirements that don't need extensive clarification
  • No breaking changes or API modifications
  • No cross-cutting concerns (security, performance, reliability)

Use full spec workflow (continue to Phase 2) when ANY of these apply:

  • Estimated implementation >80 lines of code or >3 files
  • Affects multiple subsystems or architectural boundaries
  • Requires breaking changes or significant API modifications
  • Impacts backward compatibility
  • Involves complex business logic or multiple user workflows
  • Requires coordination across multiple components
  • Has significant security, performance, or reliability implications
  • Requirements are ambiguous or need extensive clarification

When uncertain, default to the full spec workflow.

Process:

  1. Conduct initial codebase research
  2. Present sizing assessment with metrics (estimated LOC, file count, complexity factors)
  3. Recommend either smolspec or full spec workflow
  4. Get user approval for the recommended path
  5. If a Transit ticket is tracked, move it to spec status via mcp__transit__update_task_status
  6. If smolspec approved: run /starwave:smolspec, then continue to Phase 5 (Branch Creation)
  7. If full spec approved: continue to Phase 2

Phase 2: Requirement Gathering

Run the /starwave:requirements skill


Phase 3: Design Creation

Run the /starwave:design skill


Phase 4: Task Planning

Run the /starwave:tasks skill


Phase 5: Branch Creation

After tasks are approved (or after smolspec completion), offer to create a feature branch.

Process:

  1. Use the AskUserQuestion tool to offer branch naming options
  2. Include these options:

- T-{number}/{spec-name} - When a Transit ticket is tracked (recommended) - feature/{spec-name} - Standard feature branch - specs/{spec-name} - Spec-focused branch - {ticket-number}/{spec-name} - If a non-Transit ticket number was mentioned (e.g., ABC-123) - Allow user to provide a custom branch name

  1. If the user approves, create the branch and switch to it
  2. If the user declines, skip branch creation
  3. If a Transit ticket is tracked, move it to ready-for-implementation status via mcp__transit__update_task_status

Note: Only offer ticket-based branch names if a ticket was explicitly mentioned during the conversation. The Transit option should be listed first (as recommended) when a Transit ticket is present.


Response Format

Throughout the Workflow

  1. Explain what you're doing at each step
  2. Show your work (documents created, questions asked)
  3. Present review findings clearly
  4. Ask explicit approval questions
  5. Confirm what was accomplished before moving to next phase

When Gaps Are Identified

If you find gaps during any phase:

  • Mention them clearly
  • Propose relevant changes to requirements/design
  • Get user approval for changes
  • Update affected documents

User Question Handling

  • Use AskUserQuestion tool for options and choices
  • Keep questions focused and specific
  • Wait for answers before proceeding
  • Document answers in decision_log.md

Best Practices

  1. Explicit Approval Gates: Never skip approval between phases
  2. Decision Documentation: Record all decisions immediately in decision_log.md
  3. Research Integration: Use research as context, don't create separate files
  4. Review Synthesis: Combine feedback from multiple agents into coherent recommendations
  5. Incremental Refinement: Iterate with user until each phase is solid
  6. Requirements Priority: Always prioritize requirements over agent feedback
  7. Coding Focus: Tasks phase must only include coding activities
  8. Test-Driven: Emphasize testing throughout task planning

平台分布

Codex

35.05%
按下载量换算40

Claude

28.73%
按下载量换算33

Cursor

18.99%
按下载量换算22

Gemini CLI

10%
按下载量换算11

安全审计

暂无安全审计结果可展示。

权限和风险

调用外部服务

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

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