说明：本 Skill 由 阿里云（Alibaba Cloud） 提供的安全防护检查能力，用于对 OpenClaw 环境与 Skill 进行安全评估与风险检测。

---

aliyun-clawscan

Comprehensive OpenClaw security assessment combining configuration audit and skill security analysis.

---

When to Use

Scenario	Example Triggers
Full Assessment	INLINECODE0, 风险评估, INLINECODE2
Config Audit

配置审计, 检查OpenClaw配置 |
| Skill Scan | skill安全扫描, 检查skill风险 |
| Pre-install Check | 安装前检查, 这个skill安全吗 |

---

Assessment Workflow

CODEBLOCK0

---

Step 1: Configuration Audit

Run OpenClaw built-in security audit:

CODEBLOCK1

Parse results into categories (Gateway, Network, Tools, Browser, Files, Room).

Reference: reference/baseline.md for detailed check categories and parsing rules.

---

Step 2: Skill Security Audit

Phase 1: Inventory

CODEBLOCK2

Phase 2: Static Analysis

Apply local detection rules across 11 categories:

Category	Severity	Reference
Reverse Shell / Backdoor	🚨 Critical	skillaudit.md Scenario 1
Credential Harvesting

🚨 Critical | skillaudit.md Scenario 2 |
| Data Exfiltration | 🔴 High | skillaudit.md Scenario 3 |
| Cryptominer | 🚨 Critical | skillaudit.md Scenario 4 |
| Permission Abuse | 🔴 High | skillaudit.md Scenario 5 |
| Prompt Injection | 🔴 High | skillaudit.md Scenario 6 |
| Code Obfuscation | 🟡 Medium | skillaudit.md Scenario 7 |
| Ransomware | 🚨 Critical | skillaudit.md Scenario 8 |
| Persistence | 🟡 Medium | skillaudit.md Scenario 9 |
| Supply Chain | 🟡 Medium | skillaudit.md Scenario 10 |
| Malicious Service Downloader | 🚨 Critical | skillaudit.md Scenario 11 |

Reference: reference/skillaudit.md for complete detection patterns, code examples, and risk assessment logic.

Phase 3: Risk Classification

Level	Criteria
🚨 Critical	Backdoor, credential theft, ransomware, miner
🔴 High

Permission abuse, data exfil, privacy violation | | 🟡 Medium | High permissions justified, benign obfuscation | | 🟢 Low | Matches declared purpose |

---

Step 3: Consolidated Report

Report Header

CODEBLOCK3

Section 1: Configuration Audit Results

Status	Item	Finding
✅/⚠️/🔴	{Category}	{Description}

Section 2: Skill Security Findings

Risk	Count	Skills
🚨 Critical	{N}	{names}
🔴 High

{N} | {names} | | 🟡 Medium | {N} | {names} | | 🟢 Low | {N} | (see safe list) |

---

Output Templates

Quick Verdicts

Result	Message
All Clear	✅ OpenClaw风险评估完成。配置审计通过，Skill安全检查未发现明显风险。
Config Issues

⚠️ 发现配置风险。建议检查Gateway设置和文件权限配置。 | | Skill Risks | 🔴 发现Skill安全风险。{N}个高风险Skill建议立即处理。 | | Critical | 🚨 检测到严重安全风险！建议立即处理配置问题并移除恶意Skill。 |

Single Skill Assessment

• - Safe: INLINECODE11
• Sensitive: INLINECODE12
• Risky: INLINECODE13
• Malicious: INLINECODE14
• Malicious Downloader: INLINECODE15

---

Safety Guardrails

• - ✅ Static analysis only - never execute suspicious code
• ✅ Read code via file tools - no dynamic testing
• ✅ Evidence-based reporting - file:line references required
• ❌ Never bypass sandbox for analysis
• ❌ Never enumerate actual user files
• ❌ No speculation without clear evidence

---

File Reference

File	Purpose
INLINECODE16	Configuration audit rules and parsing logic
INLINECODE17

Skill security detection patterns and scenarios |

---

说明：本 Skill 由 阿里云（Alibaba Cloud） 提供的安全防护检查能力。使用本能力可对 OpenClaw 配置与已安装/待安装 Skill 进行安全审计与风险评估。