Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问许可证需确认审计提醒

find-agents寻找 Agent 商

Agent Skill

find-agents 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。

总安装

1,123

周安装

45

GitHub Stars

公开资料未说明

下载量

364
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:find-agents(寻找 Agent 商)
来源仓库:https://github.com/ulpi-io/skills
仓库路径:skills/find-agents
安装命令:
npx skills add https://github.com/ulpi-io/skills --skill find-agents
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/ulpi-io/skills --skill find-agents

简介

用于查找、检索和筛选相关信息。find-agents 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

  • 适合根据关键词或任务场景快速定位候选结果。
  • 可结合来源仓库和原始 README 核验具体用法。
  • 安装前建议确认权限范围和维护状态。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。
  • 注意是否会触发联网或文件读写操作。

SKILL.md

Non-negotiable rules:

  1. Verify agentshq availability before relying on it.
  2. Be explicit about scope: project vs global, specific IDEs vs auto-detect.
  3. Do not install or remove agents the user did not ask for.
  4. Confirm ambiguous install targets before making changes.
  5. Report exactly what was added, removed, or updated.

Find Agents

Inputs

  • $request: Search terms, repo source, or management intent such as list, remove, or update

Goal

Use agentshq to help the user:

  • discover relevant agents
  • install agents into one or more CLIs or IDEs
  • list current inventory
  • remove or update installed agents
  • scaffold a new agent when asked

Step 0: Verify runtime prerequisites

Check:

  • Node.js availability
  • npx agentshq --version

If the CLI is unavailable, explain the blocker and stop.

Success criteria: agentshq can be invoked from the current environment.

Step 1: Resolve the user's intent

Determine whether the user wants to:

  • search
  • install
  • list
  • remove
  • update
  • scaffold

Also resolve:

  • project vs global scope
  • target IDEs/CLIs if specified
  • source repo, URL, or local path if installing

If those details are ambiguous, ask before mutating anything.

Success criteria: The exact agentshq action is clear before execution.

Step 2: Run the narrowest relevant command

Command Reference

ActionCommandKey Flags
Searchnpx agentshq find [query]interactive picker when no query given
Install from reponpx agentshq add <owner>/<repo>@<agent-name> for specific agent
Install from URLnpx agentshq add <url>also works with local paths
Install scope-g for global, --ide <name>... for specific IDEs
Install options--all for all agents to all IDEs, --list to preview
List installednpx agentshq list-g for global, --ide <name>, --json
Removenpx agentshq remove [name]-g for global, --all for all agents
Check updatesnpx agentshq check
Apply updatesnpx agentshq update
Scaffold newnpx agentshq init [name]

When installing, respect:

  • explicit --ide... choices when the user gave them
  • -g only when the user asked for global scope
  • --list when the user wants to inspect a source before installing

Supported IDEs

The CLI supports 43+ IDEs. Common ones:

IDEFormat
Claude Code.md with YAML frontmatter
CursorAGENTS.md sections
GitHub Copilot.agent.md
WindsurfAGENTS.md sections
CodexAGENTS.md sections
Gemini CLI.md with YAML frontmatter
Kiro.json
AmpAGENTS.md sections
Roo CodeAGENTS.md sections

Agents are automatically translated to each IDE's native format on install.

Success criteria: The command matches the user's intent and scope exactly.

Step 3: Report the result clearly

After execution, report:

  • what command ran
  • which agents matched or changed
  • which clients/IDEs were targeted
  • any next steps such as authentication, reload, or restart

When searching, present the most relevant results instead of dumping raw output.

Success criteria: The user can immediately decide what to do next.

Guardrails

  • Do not run this skill proactively.
  • Do not install globally unless the user asked for global scope.
  • Do not remove or update agents without explicit user intent.
  • Do not guess target IDE names when the user clearly cares about a specific client.
  • Remember that agentshq can translate one source agent into multiple client formats; report that clearly.

Output Contract

Report:

  1. the resolved action
  2. project/global scope
  3. target IDEs or CLIs
  4. the agents found or changed
  5. any follow-up steps the user should know about

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

33.44%
按下载量换算122

Claude

28.56%
按下载量换算104

Cursor

19.68%
按下载量换算72

Gemini CLI

9.92%
按下载量换算36

安全审计

Gen Agent Trust Hub

通过

Socket

可疑

Snyk

可疑

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills