valkryhx/google_adk_agent / skills/search_exp
search_experience
Verified检索本地经验库,获取历史 Agent 解决过的报错方案和配置经验。
SKILL.md
View on GitHub---
name: "search_experience"
description: "检索本地经验库,获取历史 Agent 解决过的报错方案和配置经验。"
---
# Execution Instructions
## 核心能力
此技能让你拥有访问【本地集体智慧库】的能力。
这个库由 `agent_experiences/` 目录驱动,自动收录了历史 Agent 在执行任务时踩过的坑和经过验证的解决方案。通过分层索引检索,可以在几毫秒内找到最相关的历史经验。
## 工具说明
### `search_experience(query)`
**使用原则:遇到报错必须第一时间调用,不要先尝试靠猜测修复。**
| 参数 | 说明 |
| ------- | ---------------------------------------- |
| `query` | 完整的报错信息,或问题描述。越精确越好。 |
**返回值结构**:
- 【历史问题】:Agent 当时想做什么,在哪里卡住了
- 【报错特征】:原始报错信息摘要
- 【已验证的解决方案】:JSON 格式的命令/代码
- 【原理分析】:为什么会出错,为什么这样修复有效
## 使用时机
1. **遇到报错时(MUST)**:执行代码出现 Traceback、Error Code、Permission Denied 等,第一步就是调用此工具
2. **环境配置时**:需要配置 git proxy、docker、内网工具等不熟悉的命令时
3. **重复失败时**:已经尝试了 2 次以上仍然失败,查库看有没有历史解法
## 最佳实践
1. **把完整报错粘进去**:`search_experience("SSL certificate problem: unable to get local issuer certificate")` 比 `search_experience("git 报错")` 好 10 倍
2. **理解后适配**:库里的路径、参数名可能与当前任务不同,请阅读【原理分析】后自行适配,不要生搬硬套
3. **置信度判断**:返回的置信度分数越高(100 分为最高),说明与历史问题的匹配程度越精确
## 检索原理(分层设计)
```
L0 索引层 (< 1ms)
index_manifest.json <-- 只有 KB 级别,常驻检索
|
+-- 报错特征精确匹配 权重 100
+-- 标题语义相似 权重 50
+-- 关键词交集 权重 10/个
命中 & 置信度 >= 15?
|
YES --> L2 懒加载层
直接读取 agent_experiences/{category}/{gene_id}.json
返回完整解决方案
|
NO --> 返回"未找到"提示,自行分析
```
## 示例
### 示例 1:精准匹配报错
**场景**:执行脚本报错 `[ERROR] 命令执行超时(超过 60 秒)`
```python
search_experience("[ERROR] 命令执行超时(超过 60 秒)")
```
**预期输出**:
```
[search_exp] 已匹配到历史经验 (置信度: 100)
分类: python | 时间: 2026-02-22 | ID: gene_b03d793e
...
【已验证的解决方案】:
["python script.py arg1", "python script.py arg2"]
【原理分析】: 一开始使用 echo 管道输入交互模式,但 input() 无法处理
管道输入导致超时。后来改为命令行参数调用绕过了交互模式。
```
### 示例 2:关键词语义匹配
**场景**:不确定当前环境 git 代理如何配置
```python
search_experience("git proxy 配置 内网 ssl error")
```
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| 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: 472k | Updated | Claude CodeCodex | Verified | Install |
| 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: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| 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 available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |