针对 OpenClaw 技能的综合安全套件。包括静态扫描(AST + 关键字)和人工智能驱动的语义行为审查以检测恶意代码。
Install to Claude Code
npx -y skills add lza6/Claude-code-cli-config --skill openclaw-security-suite-0.2.3 --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Openclaw Security Suite 0.2.3?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/lza6-openclaw-security-suite-0-2-3)More formats (shields.io, HTML) on the badges page.
---
name: openclaw-security-suite
description: "针对 OpenClaw 技能的综合安全套件。包括静态扫描(AST + 关键字)和人工智能驱动的语义行为审查以检测恶意代码。"
version: 0.2.3
tags:
- security
- scanner
- code-review
- static-analysis
- ai-review
requirements:
binaries:
- node
environment: []
input_schema:
type: object
properties:
action:
type: string
enum: ["scan", "review"]
description: 执行的操作。"scan" 用于静态分析,"review" 用于 AI 语义审查。
path:
type: string
description: 技能目录(扫描)或文件(审查)的绝对路径。
required:
- action
- path
---
# OpenClaw 安全套件 (OpenClaw Security Suite)
OpenClaw 扩展的全面安全保护层,提供静态分析和 AI 辅助行为审查。
## 特性
该套件整合了两个核心安全功能:
### 1. 静态安全扫描 (`action: "scan"`)
对整个技能目录进行分析,以识别确定的威胁:
- **已屏蔽的导入**:例如 `child_process`、`cluster`
- **危险函数**:例如 `exec()`、`spawn()`
- **已知的恶意关键字**:例如 `eval(`、`__proto__`、`rm -rf`
- **敏感文件访问**:例如 `/etc/passwd`、`/.env`
- **可疑的正则模式**:例如 `curl ... | bash`
### 2. AI 代码审查 (`action: "review"`)
利用活跃的 LLM 上下文 (`ctx.llm`) 对特定文件进行语义分析,查找隐藏威胁:
- **数据外泄**
- **凭据泄露**
- **混淆后的 Shell 执行**
- **系统篡改**
## 用法
您必须指定 `action` 和 `path` 参数。
**示例 1:静态扫描**
```json
{
"action": "scan",
"path": "/path/to/skill/directory"
}
```
**示例 2:AI 审查**
```json
{
"action": "review",
"path": "/path/to/skill/index.ts"
}
```
## 输出示例
**扫描输出:**
```json
{
"safe": false,
"results": [
{
"file": "index.ts",
"issues": [{ "type": "blocked_module", "module": "child_process" }]
}
]
}
```
**审查输出:**
```json
{
"risk_level": "high",
"reason": "代码从环境变量读取 AWS 凭据并将其发送至外部 IP 地址。"
}
```
Scanned 9/12/2026
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!