中文 EN

扫描报告

扫描新文件

⚠ 低风险 — 风险评分 20/100
上次扫描：1 天前 重新扫描
20 /100
enterprise-memory
EAM规范企业级Agent记忆体系Skill，提供项目级隔离、状态同步、快照压缩及SOP流程管理能力
Legitimate EAM project management skill with well-structured SOP workflows. Minor documentation gap regarding network access for web fetching capability.
技能名称enterprise-memory
分析耗时40.0s
引擎pi
✓
可以安装
Consider adding explicit network:READ capability declaration in SKILL.md for the jina.ai web fetching feature. The optional LLM compression warning is appropriate.

安全发现 2 项

严重性 安全发现 位置
低危
Network capability not mapped in capability model 文档欺骗
The skill documents jina.ai web fetching capability using curl, but does not include network:READ in the allowed-tools mapping. This is a documentation gap rather than hidden functionality.
curl -s "https://r.jina.ai/[URL]"
→ Add network:READ to capability model if web fetching is part of the skill's core functionality
 SKILL.md:327
提示
Subprocess calls between scripts 代码执行
Scripts call each other via subprocess.run() for state synchronization. This is legitimate internal coordination and documented.
subprocess.run([sys.executable, str(update_script), '--project-dir', ...])
→ No action needed - this is expected behavior for script orchestration
 scripts/sop_state.py:90
资源类型声明权限推断权限状态证据
文件系统 READ,WRITE READ,WRITE ✓ 一致 SKILL.md:89-100 defines ~/.openclaw/EAM-projects/ storage
命令执行 WRITE WRITE ✓ 一致 subprocess.run() calls between scripts for internal coordination
网络访问 NONE READ ✓ 一致 curl to jina.ai in SKILL.md:327-365 but not in capability mapping
3 项发现
🔗
中危 外部 URL 外部 URL
https://clawhub.ai/evan-zhang/enterprise-memory
SKILL.md:18
🔗
中危 外部 URL 外部 URL
https://r.jina.ai/[URL
SKILL.md:327
🔗
中危 外部 URL 外部 URL
https://r.jina.ai/https://toutiao.com/article/12345
SKILL.md:363

目录结构

26 文件 · 117.9 KB · 3770 行
Python 8f · 2427L Markdown 16f · 1310L JSON 2f · 33L
├─ ▾ 📁 references
│ ├─ ▾ 📁 shared
│ │ ├─ 📝 confirm-protocol.md Markdown 38L · 1.4 KB
│ │ ├─ 📝 state-machine.md Markdown 55L · 1.5 KB
│ │ └─ 📝 upgrade-rules.md Markdown 53L · 1.5 KB
│ ├─ ▾ 📁 templates
│ │ ├─ ▾ 📁 full
│ │ │ ├─ 📝 ARTIFACTS-template.md Markdown 33L · 710 B
│ │ │ ├─ 📝 DECISIONS-template.md Markdown 27L · 735 B
│ │ │ └─ 📝 PLAN-template.md Markdown 40L · 751 B
│ │ ├─ ▾ 📁 lite
│ │ │ ├─ 📝 HANDOVER-template.md Markdown 45L · 768 B
│ │ │ ├─ 📝 LOG-template.md Markdown 37L · 780 B
│ │ │ ├─ 📝 RESULT-template.md Markdown 54L · 912 B
│ │ │ └─ 📝 TASK-template.md Markdown 52L · 821 B
│ │ ├─ 📝 DECISIONS.md Markdown 9L · 173 B
│ │ ├─ 📝 INDEX.md Markdown 31L · 451 B
│ │ └─ 📋 state.json JSON 28L · 593 B
│ ├─ 📝 sop-full-guide.md Markdown 107L · 3.4 KB
│ └─ 📝 sop-lite-guide.md Markdown 91L · 2.8 KB
├─ ▾ 📁 scripts
│ ├─ 🐍 __init__.py Python 2L · 63 B
│ ├─ 🐍 compress.py Python 429L · 14.3 KB
│ ├─ 🐍 sop_handover.py Python 248L · 8.1 KB
│ ├─ 🐍 sop_init.py Python 283L · 9.7 KB
│ ├─ 🐍 sop_state.py Python 286L · 11.5 KB
│ ├─ 🐍 sop_upgrade.py Python 393L · 14.0 KB
│ ├─ 🐍 switch_project.py Python 489L · 17.3 KB
│ └─ 🐍 update_index.py Python 297L · 8.8 KB
├─ 📋 _meta.json JSON 5L · 136 B
├─ 📝 CHARTER.md Markdown 230L · 4.6 KB
└─ 📝 SKILL.md Markdown 408L · 12.2 KB

安全亮点

✓ All file operations use atomic writes (temp file → rename pattern)
✓ Proper security warning in SKILL.md for optional LLM compression feature
✓ No credential theft or sensitive file access (~/.ssh, ~/.aws, .env)
✓ No obfuscation or base64-encoded payloads
✓ No external dependencies beyond Python standard library
✓ Well-structured state management with schema validation
✓ Dry-run support for all destructive operations
✓ High-risk operations require explicit --confirm flag