Skip to content

valkryhx/google_adk_agent / skills/use_external_skills

use_external_skills

Verified

从外部 skills 库(~/.agents/skills)中查找、匹配并加载最合适的 skill 指令。注意:本 skill 专门用于操作外部 skill 库,加载本项目内置 skill 请使用 skill_load

FileKindSize
SKILL.mdskill md2.5 KB
tools.pyscript11.4 KB
---
name: use_external_skills
description: 从外部 skills 库(~/.agents/skills)中查找、匹配并加载最合适的 skill 指令。注意:本 skill 专门用于操作外部 skill 库,加载本项目内置 skill 请使用 skill_load
---

# Use External Skills

> **重要区分**
> - 加载**本项目内置** skill(`skills/` 目录)→ 使用 `skill_load <id>`
> - 加载**外部** skill(`C:\Users\drago\.agents\skills`)→ 使用本 skill 提供的工具

本 skill 从 `C:\Users\drago\.agents\skills` 目录中动态发现并加载外部 skill 的 Instructions 及脚本资产。

## 工具一览

| 工具名 | 用途 |
|--------|------|
| `list_external_skills()` | 列出全部可用外部 skill 及其描述 |
| `search_external_skill(query)` | 按意图关键词搜索最匹配的外部 skill(返回 top-3)|
| `load_external_skill(skill_id)` | 加载指定外部 skill 的完整 Instructions + 所有附加 .md 文档 |
| `list_external_skill_assets(skill_id)` | 列出该外部 skill 目录下的全部资产(脚本/文档/引用文件)|
| `run_external_skill_script(skill_id, script_path, args)` | 执行该外部 skill 中的脚本(支持 .sh/.js/.ts/.py)|

## 标准使用流程

1. **发现**:调用 `search_external_skill(query)` 描述你的意图,获取推荐 skill
2. **了解**:调用 `load_external_skill(skill_id)` 获取完整 Instructions,**加载后必须严格遵循其中指令**
3. **查看资产**(可选):调用 `list_external_skill_assets(skill_id)` 查看是否有可执行脚本
4. **执行脚本**(可选):调用 `run_external_skill_script(skill_id, "scripts/start-server.sh")` 运行脚本

## 示例

```
# 查找调试相关的外部 skill
search_external_skill("调试代码 bug 根因分析")
→ 推荐: systematic-debugging

# 加载并遵循其 Instructions
load_external_skill("systematic-debugging")

# 查看 brainstorming 的脚本资产
list_external_skill_assets("brainstorming")
→ [script] scripts/start-server.sh
→ [script] scripts/server.js

# 启动 brainstorming 预览服务器
run_external_skill_script("brainstorming", "scripts/start-server.sh", "--project-dir /tmp/demo")
```

## 注意事项

- 外部 skill 加载后,其 Instructions 具有与本项目内置 skill 同等的执行优先级
- 若外部 skill 的 Instructions 要求执行 shell 命令,通过已加载的 `bash` skill 工具完成
- 若无完全匹配,`search_external_skill` 会返回相关度最高的候选,由 agent 自行判断是否合适
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall