扫描报告
20 /100
zhua-evolver
爪爪专属自我进化系统 —— 自动分析能力差距、搜索补强技能、执行进化循环、记录进化日志
This is a self-evaluation benchmarking skill for '爪爪' AI agent with no malicious capabilities. Minor documentation discrepancies exist but no security-impacting issues found.
可以安装
Approve for use. Documented scripts (search_skills.py, install_skills.py, log_evolution.py) should be implemented or removed from SKILL.md to eliminate documentation mismatch.
安全发现 2 项
| 严重性 | 安全发现 | 位置 |
|---|---|---|
| 低危 | Documented capabilities not implemented 文档欺骗 | SKILL.md:1 |
| 低危 | Referenced scripts do not exist 文档欺骗 | SKILL.md:25 |
| 资源类型 | 声明权限 | 推断权限 | 状态 | 证据 |
|---|---|---|---|---|
| 文件系统 | NONE | NONE | — | No file operations found in any scripts |
| 网络访问 | READ | NONE | ✗ 越权 | SKILL.md claims skillhub searching capability but no network code exists in scri… |
| 命令执行 | NONE | NONE | — | No subprocess or shell execution found |
| 环境变量 | NONE | NONE | — | No os.environ access or credential harvesting |
| 剪贴板 | NONE | NONE | — | No clipboard access found |
目录结构
10 文件 · 21.3 KB · 694 行 Python 7f · 561L
Markdown 2f · 109L
Text 1f · 24L
├─
▾
assets
│ └─
example_asset.txt
Text
├─
▾
references
│ └─
api_reference.md
Markdown
├─
▾
scripts
│ ├─
analyze_gap.py
Python
│ ├─
check_hyper_top_final_v2.py
Python
│ ├─
check_hyper_top_final.py
Python
│ ├─
check_hyper_top_v2.py
Python
│ ├─
define_hyper_top.py
Python
│ ├─
example.py
Python
│ └─
orchestrate_minions.py
Python
└─
SKILL.md
Markdown
安全亮点
✓ No network access - all scripts are pure computation without network calls
✓ No credential harvesting - no environment variable access for secrets or API keys
✓ No file write operations - all scripts perform read-only computations
✓ No shell execution - pure Python scripts using argparse without subprocess
✓ No base64 obfuscation or encoded payloads
✓ No sensitive path access (~/.ssh, ~/.aws, .env, etc.)
✓ All dependencies are Python standard library only - no external package requirements
✓ No C2 communication or data exfiltration indicators
✓ No persistence mechanisms (cron, startup scripts, backdoors)
✓ No prompt injection attempts detected