中文 EN

扫描报告

扫描新文件

⚠ 低风险 — 风险评分 15/100
上次扫描：1 天前 重新扫描
15 /100
lmail_ops_complete
Operate LMail end-to-end with strict registration, authentication, inbox loops, threaded replies, and admin registration audits
This is a legitimate LMail email operations skill with proper credential handling, standard library-only dependencies, and no malicious behavior. Minor concerns include undocumented credential file writes and a hardcoded example URL.
技能名称lmail_ops_complete
分析耗时65.6s
引擎pi
✓
可以安装
This skill is safe to use. Ensure the LMAIL_BASE_URL is trusted before deployment. Consider pinning the base URL in production deployments.

安全发现 3 项

严重性 安全发现 位置
低危
Undeclared credential file writes 文档欺骗
SKILL.md does not declare that scripts write credentials to .lmail-credentials.json. This is necessary functionality but should be documented.
save_json_file(args.credentials_file, creds)
→ Add credential persistence to SKILL.md documentation
 scripts/strict_register.py:158
提示
External service URL reference 敏感访问
The skill references https://amiigzz.online as an example base URL. This requires trust in the service operator.
https://amiigzz.online
→ Verify the service operator is trusted before using
 SKILL.md:56
提示
No dependency pinning 供应链
Scripts use only Python standard library, but no requirements.txt or dependency pinning exists. This is acceptable since no external packages are used.
import os, json, sys - all standard library
→ No action needed - standard library only
 N/A
资源类型声明权限推断权限状态证据
文件系统 NONE WRITE ✓ 一致 Writes credentials.json and state files - necessary for the feature
网络访问 READ READ ✓ 一致 All network calls to configured LMAIL_BASE_URL
命令执行 NONE WRITE ✓ 一致 Shell scripts for deployment - documented in SKILL.md
3 项发现
🔗
中危 外部 URL 外部 URL
https://amiigzz.online
SKILL.md:56
📧
提示 邮箱 邮箱地址
[email protected]
SKILL.md:115
📧
提示 邮箱 邮箱地址
[email protected]
assets/payload-templates/send-message.json:2

目录结构

31 文件 · 70.8 KB · 2387 行
Python 11f · 1467L Shell 8f · 511L Markdown 8f · 382L JSON 3f · 23L YAML 1f · 4L
├─ ▾ 📁 agents
│ └─ 📋 openai.yaml YAML 4L · 252 B
├─ ▾ 📁 assets
│ ├─ ▾ 📁 curl-snippets
│ │ └─ 🔧 challenge-solve-register.sh Shell 10L · 312 B
│ └─ ▾ 📁 payload-templates
│ ├─ 📋 override-permit.json JSON 3L · 64 B
│ ├─ 📋 register.json JSON 8L · 205 B
│ └─ 📋 send-message.json JSON 12L · 215 B
├─ ▾ 📁 references
│ ├─ 📝 api-contract.md Markdown 37L · 880 B
│ ├─ 📝 docs-source-of-truth.md Markdown 26L · 662 B
│ ├─ 📝 error-codes-runbook.md Markdown 33L · 1.0 KB
│ ├─ 📝 openclaw-publish-checklist.md Markdown 62L · 1.4 KB
│ ├─ 📝 ops-playbook.md Markdown 30L · 999 B
│ ├─ 📝 security-policy.md Markdown 21L · 720 B
│ └─ 📝 strict-registration-v2.md Markdown 43L · 1.3 KB
├─ ▾ 📁 scripts
│ ├─ 🐍 admin_fetch_registration_events.py Python 85L · 2.6 KB
│ ├─ 🐍 admin_issue_override_permit.py Python 87L · 2.7 KB
│ ├─ 🐍 chat_fast.py Python 261L · 8.4 KB
│ ├─ 🐍 inbox_loop.py Python 109L · 3.7 KB
│ ├─ 🐍 inbox_once.py Python 168L · 5.8 KB
│ ├─ 🔧 install_for_openclaw.sh Shell 49L · 994 B
│ ├─ 🐍 lmail_http.py Python 147L · 4.2 KB
│ ├─ 🐍 login_verify.py Python 114L · 3.4 KB
│ ├─ 🔧 preflight_check.sh Shell 49L · 1.0 KB
│ ├─ 🔧 print_claude_setup.sh Shell 21L · 680 B
│ ├─ 🔧 publish_clawhub.sh Shell 134L · 3.0 KB
│ ├─ 🐍 send_message.py Python 109L · 3.5 KB
│ ├─ 🐍 send_reply.py Python 112L · 3.4 KB
│ ├─ 🔧 smoke_test_registration_flow.sh Shell 102L · 2.7 KB
│ ├─ 🐍 solve_pow.py Python 84L · 2.4 KB
│ ├─ 🐍 strict_register.py Python 191L · 6.2 KB
│ ├─ 🔧 sync_docs_check.sh Shell 72L · 1.3 KB
│ └─ 🔧 validate_openclaw_skill.sh Shell 74L · 1.9 KB
└─ 📝 SKILL.md Markdown 130L · 5.0 KB

依赖分析 1 项

包名版本来源已知漏洞备注
python3-standard-library N/A built-in 否 No external dependencies - uses only stdlib

安全亮点

✓ Standard library only - no malicious dependencies possible
✓ Credentials are masked when printed (mask_secret function)
✓ Credentials files set to mode 0600
✓ No subprocess/eval/exec usage
✓ No base64 execution into shell - base64 only used for legitimate JWT parsing
✓ No access to ~/.ssh, ~/.aws, .env, or other sensitive paths
✓ No curl|bash or wget|sh remote script execution
✓ All network requests go to configured base URL, no hardcoded IPs
✓ Well-structured code with proper error handling
✓ Security policy document present in references/