Token导航 LogoToken导航

Agent Skill · 待分类

clarify(文案澄清)

clarify 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。

clarify 是待分类类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;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 8,658。

总安装 285周安装 12GitHub Stars 公开资料未说明下载 1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:clarify(文案澄清)
来源仓库:https://github.com/vitadynamics/vita-cc-market
仓库路径:skills/clarify
安装命令:
npx skills add https://github.com/vitadynamics/vita-cc-market --skill clarify
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/vitadynamics/vita-cc-market --skill clarify

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。

用于处理 GitHub 仓库、Issue 和 Pull Request 信息,适合协作事项整理。可帮助 Agent 围绕代码变更和仓库状态进行梳理,提升开发流程效率。

安装命令:npx skills add https://github.com/vitadynamics/vita-cc-market --skill clarify 使用前建议确认权限范围和维护状态,避免触发联网或命令执行。注意检查是否会读写文件或修改项目结构。

来源仓库内路径:skills/clarify

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/vitadynamics/vita-cc-market --skill clarify;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

    它需要需要联网,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

clarify 怎么安装?

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

clarify 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。

clarify 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

clarify 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。

clarify 用的人多吗?

公开数据是累计安装 285 次,近一周安装 12 次,下载 1 次。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。

clarify 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
clarify

文案澄清 · 本页

285—需要联网未明确通过

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

Clarify Skill

Purpose: Detect and reduce ambiguity in feature requests by asking targeted clarification questions and updating plans with the results.

User Input: "${ARGUMENTS}"

If the input above is empty, ask the user using AskUserQuestion: "What feature, bug, or improvement would you like to clarify? Please provide a brief description."

Do not proceed until you have input to clarify.

Stage 1: Ambiguity & Coverage Scan

Analyze the user input against this taxonomy. For each category, mark status: Clear / Partial / Missing.

Taxonomy Categories

CategoryWhat to Check
Functional ScopeCore user goals, success criteria, out-of-scope declarations
Domain & Data ModelEntities, attributes, relationships, identity rules, state transitions
Interaction & UX FlowUser journeys, error/empty/loading states, accessibility
Non-Functional RequirementsPerformance targets, scalability, reliability, security, observability
Integration & DependenciesExternal services/APIs, data formats, failure modes
Edge Cases & ErrorsNegative scenarios, rate limiting, conflict resolution
Constraints & TradeoffsTechnical constraints, explicit tradeoffs, rejected alternatives
TerminologyCanonical terms, avoided synonyms

Coverage Assessment

After scanning, produce an internal coverage map:

  • Clear: Sufficient information provided
  • Partial: Some information but gaps exist
  • Missing: No information provided, critical for implementation

Stage 2: Generate Clarification Questions

Based on the coverage scan, generate a prioritized queue of up to 5 questions.

Question Constraints

Each question must be:

  • Answerable with multiple choice (2-5 options) OR short answer (≤5 words)
  • High impact: Materially affects architecture, data modeling, task decomposition, or testing
  • Not already answered in the user input
  • Not trivial stylistic preferences

Question Prioritization

Prioritize by: Impact × Uncertainty

  • Cover highest-impact unresolved categories first
  • Avoid asking two low-impact questions when a high-impact area is unresolved

Stage 3: Sequential Questioning (Using AskUserQuestion Tool)

IMPORTANT: Use the AskUserQuestion tool to present each question. Present ONE question at a time.

For Multiple-Choice Questions

  1. Analyze options and determine the most suitable based on:

- Best practices for the project type - Common patterns in similar implementations - Risk reduction (security, performance, maintainability)

  1. Use AskUserQuestion with options:

- Put recommended option first with "(Recommended)" suffix - Provide 2-4 clear options - Include description for each option

  1. Example AskUserQuestion call: Question: "How should user authentication be handled?" Options: - "JWT tokens (Recommended)" - Stateless, scalable, industry standard - "Session-based" - Server-side sessions with cookies - "OAuth only" - Delegate to third-party providers

For Short-Answer Questions

Use AskUserQuestion with options that include common choices:

Question: "What is the expected maximum number of concurrent users?"
Options:
- "< 100 users" - Small scale, simple infrastructure
- "100-1000 users" - Medium scale, may need caching
- "1000+ users (Recommended)" - Large scale, requires optimization

After Each Answer

  • Record the answer
  • Move to next question
  • Stop when: all critical ambiguities resolved, user signals "done", or 5 questions asked

Stage 4: Update Plan with Clarifications

After all questions are answered, append clarifications to the plan file.

If plan file exists (plans/<topic>.md):

Add a new section to the plan:

## Clarifications

**Session:** [DATE]

### Questions & Answers

| # | Question | Answer | Category |
|---|----------|--------|----------|
| 1 | [Question 1] | [Answer 1] | Functional Scope |
| 2 | [Question 2] | [Answer 2] | Data Model |
| ... | ... | ... | ... |

### Coverage Summary

| Category | Status |
|----------|--------|
| Functional Scope | Resolved |
| Domain & Data Model | Clear |
| Integration & Dependencies | Deferred |
| ... | ... |

### Key Decisions

- **[Decision 1]**: [Rationale based on answer]
- **[Decision 2]**: [Rationale based on answer]

If plan file does not exist:

Create a new clarifications file at plans/clarifications-<topic>.md with the same format, which can be referenced when running /core:plan.

Stage 5: Next Steps

Use AskUserQuestion to present options:

Question: "Clarification complete. [X] questions answered. Plan updated. What would you like to do next?"

Options:

  1. Run /core:plan - Create/update the plan with these clarifications (Recommended)
  2. Ask more questions - Continue clarifying specific areas
  3. View clarifications - Show the clarifications section that was added
  4. Start over - Reset and re-clarify from scratch

Behavior Rules

  • Always use AskUserQuestion tool for gathering user input
  • If no meaningful ambiguities found → "No critical ambiguities detected. Ready for /core:plan."
  • Never exceed 5 questions total
  • Respect early termination signals ("done", "stop", "proceed")
  • Do not ask speculative tech stack questions unless blocking functional clarity
  • Always update the plan file with clarifications after completion

平台分布

Codex

36.89%
按下载量换算0

Claude

32.26%
按下载量换算0

Cursor

19.02%
按下载量换算0

Gemini CLI

9.58%
按下载量换算0

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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