奥辛特多尔金

工程开发 社区

技能解读

当前原文的自动摘译,尚未人工审校。

为什么采用这个工作流程

利用数据库页面可以实施反机器人保护和速率限制。无需抓取 HTML 页面,而是通过捆绑的帮助程序脚本使用 ExploitDB GitLab 存储库中的官方 GHDB XML 提要。帮助程序使用的主要数据源:https://gitlab.com/exploit-database/exploitdb/-/raw/main/ghdb.xml

示例

使用这些作为一些简短的参考,了解强大、实用的呆子是什么样子的。暴露的环境文件: Dork: inurl:.env OR intitle:"index of" ".env" 为什么有用:有助于检测配置错误的部署中泄露的运行时机密。打开目录列表:

安全和范围规则

仅提供授权、合法和防御用途的潜入指导。如果用户请求明显未经授权的目标,请拒绝并提供良性替代方案(强化检查、自我审核查询、培训/实验室示例)。比起侵入性的指导,更喜欢危害最小、注重发现的呆子。不提供分步利用说明。

程序

明确意图和限制。询问用户想要检测什么技术、文件类型、产品或暴露类别。如果需要,询问他们是否想要广泛的发现或高度具体的呆子。通过捆绑脚本查询 dork 数据。从此技能目录运行命令:skills/osint-dorking。

输出模板

在响应中使用此结构:匹配 1:ID:<id> 类别:<category> Dork:<query>

流狐档案 运行与权限由流狐静态检查推断,不代表作者声明或实际测试结果。
领域
工程开发
作者声明的智能体
未找到明确声明;不据此推断已兼容或已测试
静态检查
94 / 100 · 启发式扫描,不代表运行安全
运行环境证据
未执行;当前没有隔离运行测试证据
作者 / 版本 / 许可证
@000001000000 · MIT
上下文用量评级
低消耗
接入复杂程度
需简单配置
是否需要外部接口密钥
不需要
兼容的系统
未声明(默认跨平台)
运行环境要求
Python
检测到的文件与系统行为
  • 只读
  • 命令执行
  • 读取环境变量
网络行为
允许外网请求
安装命令数
无(仅作为资料)

需要注意: 未声明工具权限范围;实际权限由运行环境决定。

作者说明文件(原文) SKILL.md

# OSINT Dorking Skill

This skill teaches an agent how to reliably answer OSINT dorking requests. It currently uses GHDB as a robust baseline data source and can be extended with additional dork sources.

## Why this workflow

Exploit-DB pages can enforce anti-bot protections and rate limits. Instead of scraping HTML pages, use the official GHDB XML feed from the ExploitDB GitLab repository through the bundled helper script.

Primary data source used by the helper:
- https://gitlab.com/exploit-database/exploitdb/-/raw/main/ghdb.xml

## Examples

Use these as few-shot references for what strong, practical dorks look like.

- Exposed environment files:
  - Dork: inurl:.env OR intitle:"index of" ".env"
  - Why useful: helps detect leaked runtime secrets in misconfigured deployments.
- Open directory listings:
  - Dork: intitle:"index of" "backup" OR intitle:"index of" "/uploads"
  - Why useful: surfaces browseable file indexes that may expose sensitive artifacts.
- WordPress configuration backups:
  - Dork: intitle:"index of" "wp-config.php.bak"
  - Why useful: identifies accidental backup exposure of database credentials.
- Database dump files:
  - Dork: ext:sql "dump" ("db" OR "backup")
  - Why useful: finds potentially exposed SQL exports for defensive verification.
- Git metadata exposure:
  - Dork: intitle:"index of" ".git"
  - Why useful: highlights repositories accidentally exposed via directory listing.

## Safety and scope rules

1. Only provide dorking guidance for authorized, legal, and defensive use.
2. If the user requests clearly unauthorized targeting, refuse and offer benign alternatives (hardening checks, self-audit queries, training/lab examples).
3. Prefer least harmful, discovery-focused dorks over intrusive guidance.
4. Do not provide step-by-step exploitation instructions.

## Procedure

1. Clarify intent and constraints.
- Ask what technology, file type, product, or exposure class the user wants to detect.
- If needed, ask whether they want broad discovery or highly specific dorks.

2. Query dork data via bundled script.
- Run commands from this skill directory: `skills/osint-dorking`.
- Run:
  - ./scripts/search.sh --keyword "<term>" --limit 10
- Equivalent direct Python entrypoint:
  - python3 scripts/search_dorks.py --keyword "<term>" --limit 10
- For category filters:
  - ./scripts/search.sh --keyword "<term>" --category "Files Containing Juicy Info" --limit 10
- For exact reproducibility in machine-readable format:
  - ./scripts/search.sh --keyword "<term>" --format json --limit 20

3. If results are weak, iterate.
- Try synonyms and related products.
- Switch between broad and narrow terms.
- Use multiple keywords with match-any mode:
  - ./scripts/search.sh --keyword "wordpress" --keyword "wp-config" --match any --limit 20

4. Respond with curated output.
- Return a concise list including:
  - GHDB ID
  - Category
  - Dork query
  - Why it is relevant
- Add a short caution reminding the user to test only systems they own or are authorized to assess.

5. If the feed cannot be fetched.
- Retry once with --refresh.
- If network still fails, rely on cached data at data/ghdb.xml.
- Tell the user the data may be stale and provide date context if available.

## Output template

Use this structure in responses:

- Match 1:
  - ID: <id>
  - Category: <category>
  - Dork: <query>
  - Relevance: <reason>

- Match 2:
  - ID: <id>
  - Category: <category>
  - Dork: <query>
  - Relevance: <reason>

## Notes for agents

- Preferred entrypoint is ./scripts/search.sh because it handles Python interpreter selection and forwards args safely.
- Prefer running the local script over direct Fetch calls to exploit-db pages.
- Keep answers practical and defensive.
- If the user asks for more, expand with adjacent categories and safer variants.

讨论

基于 GitHub Discussions。登录 GitHub 即可参与讨论、点赞、订阅更新。