Token导航 LogoToken导航

Agent Skill · 研究检索

tool-plan-interviewer(工具计划面试官)

tool-plan-interviewer 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。

tool-plan-interviewer 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,按安装量在研究检索类的 73,675 个 Skill 里排第 62,244。

总安装 339周安装 14GitHub Stars 2下载 111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:tool-plan-interviewer(工具计划面试官)
来源仓库:https://github.com/hungv47/agent-skills
仓库路径:skills/tool-plan-interviewer
安装命令:
npx skills add https://github.com/hungv47/agent-skills --skill tool-plan-interviewer
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/hungv47/agent-skills --skill tool-plan-interviewer

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。

tool-plan-interviewer 用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据任务场景快速定位候选结果。通过 npx skills add 命令从指定 GitHub 仓库安装使用。安装前需确认权限范围、维护状态,以及是否涉及联网或文件操作。

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/tool-plan-interviewer

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/hungv47/agent-skills --skill tool-plan-intervi…;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

tool-plan-interviewer 怎么安装?

在终端执行 npx skills add https://github.com/hungv47/agent-skills --skill tool-plan-interviewer。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

tool-plan-interviewer 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。

tool-plan-interviewer 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。

tool-plan-interviewer 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。

tool-plan-interviewer 用的人多吗?

公开数据是累计安装 339 次,近一周安装 14 次,下载 111 次,GitHub 2 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。

tool-plan-interviewer 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。
  • 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
tool-plan-interviewer

工具计划面试官 · 本页

3392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

Plan Interviewer

Conduct comprehensive, multi-round interviews when entering plan mode to ensure thorough understanding before implementation.

Purpose

Transform shallow feature requests into well-specified plans by asking probing, non-obvious questions that surface hidden complexity, edge cases, architectural decisions, and user experience considerations.

When to Use

  • Upon entering plan mode for any non-trivial task
  • When a spec file or plan file is mentioned
  • Before writing or updating implementation specifications
  • When the user asks to "think through" or "help me plan" something

Core Process

Phase 1: Context Gathering

  1. Read existing artifacts: If a plan file or spec file exists (or is mentioned), read it first to understand current state
  2. Identify the domain: Determine whether this is frontend, backend, full-stack, infrastructure, data, etc.
  3. Assess complexity: Estimate how many rounds of questioning are needed

Phase 2: Multi-Round Interview

Conduct interviews using AskUserQuestion tool. Each round should explore different dimensions. Never ask obvious questions like "what framework?" or "what language?" when context already provides this.

Interview Dimensions (cycle through these):

Technical Implementation
  • Data flow edge cases (What happens when X fails mid-operation?)
  • State management boundaries (Where does this state live? Who owns it?)
  • Concurrency considerations (What if two users do this simultaneously?)
  • Migration paths (How do we get from current state to this without breaking things?)
  • Performance cliffs (At what scale does this approach break down?)
UX & Interaction Design
  • Error recovery flows (User does X wrong—then what?)
  • Loading/empty/error states (What does the user see during transitions?)
  • Undo/redo expectations (Should this be reversible? How far back?)
  • Accessibility edge cases (How does a screen reader announce this state change?)
  • Mobile-first considerations (Does this gesture conflict with system gestures?)
Business Logic & Domain
  • Edge case ownership (Who decides what happens when rules conflict?)
  • Audit/compliance needs (Does this action need to be logged? For how long?)
  • Multi-tenancy implications (How does this behave across different accounts/orgs?)
  • Internationalization (Does this text expand 40% in German? RTL support?)
Architecture & Tradeoffs
  • Consistency vs availability (What's acceptable during network partition?)
  • Coupling concerns (Does this create a dependency we'll regret?)
  • Testability (How do we verify this works without manual testing?)
  • Observability (How do we know this is broken in production?)
Security & Privacy
  • Trust boundaries (What if this input comes from an attacker?)
  • Data exposure (What PII touches this flow? Who can see it?)
  • Rate limiting (What happens if someone calls this 10,000 times?)

Phase 3: Synthesis & Specification

After sufficient interview rounds (typically 3-7 depending on complexity):

  1. Summarize findings: Recap the key decisions made during interview
  2. Write specification: Create or update the spec/plan file with:

- Clear problem statement - Decided approach with rationale - Edge cases and their handling - Open questions (if any remain) - Implementation notes

Interview Technique Guidelines

Question Quality Rules

Avoid obvious questions:

  • ❌ "What tech stack?" (look at the codebase)
  • ❌ "What should the button say?" (context usually provides this)
  • ❌ "Should we use TypeScript?" (check existing code)

Ask revealing questions:

  • ✅ "If the webhook fails after the database write succeeds, should we retry, rollback, or alert?"
  • ✅ "When the cache is stale but the origin is down, do we serve stale or error?"
  • ✅ "If a user is mid-edit and their session expires, do we save their work or lose it?"

Question Framing

Frame questions to surface hidden assumptions:

  • "What's the worst thing that could happen if..."
  • "When you say X, do you mean A or B? Because they have different implications for..."
  • "I notice the current system does Y—should we preserve that behavior or is this a chance to fix it?"
  • "Who's the angriest user for this feature and what would make them happy?"

Pacing

  • Ask 2-4 questions per round maximum
  • Group related questions together
  • After each round, briefly acknowledge answers before next questions
  • Stop when answers start repeating or user indicates completion

AskUserQuestion Tool Usage

Structure questions for maximum clarity:

Use AskUserQuestion with:
- 2-4 questions per invocation
- Each question has 2-4 concrete options
- Options should represent real tradeoffs, not obvious choices
- Include brief description explaining implications of each option

Example question structure:

Question: "When a background sync fails, how should we handle it?"
Options:
1. Silent retry (3x with backoff) - User unaware, but may see stale data
2. Toast notification - User informed but may be annoyed
3. Badge indicator - Subtle, user can investigate when ready
4. Block UI until resolved - Safest but worst UX

Spec File Format

When writing the specification, use this structure:

# [Feature Name] Specification

## Problem Statement
[1-2 paragraphs on what we're solving and why]

## Decided Approach
[High-level approach with key architectural decisions]

## Key Decisions
| Decision | Choice | Rationale |
|----------|--------|-----------|
| [Topic]  | [What] | [Why]     |

## Edge Cases
- **[Scenario]**: [How we handle it]

## Out of Scope
- [What we're explicitly NOT doing]

## Open Questions
- [ ] [Anything still unresolved]

## Implementation Notes
[Technical details, gotchas, dependencies]

Completion Criteria

Stop interviewing when:

  • All major dimensions have been covered
  • User indicates readiness to proceed
  • Answers are becoming repetitive
  • Core architecture and edge cases are documented

Then write the spec to the designated file and confirm with user.

Additional Resources

Reference Files

  • references/question-bank.md - Extended list of probing questions by domain

平台分布

Codex

35.68%
按下载量换算40

Claude

31.17%
按下载量换算35

Cursor

18.65%
按下载量换算21

Gemini CLI

10.2%
按下载量换算11

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