可信 — 风险评分 5/100
上次扫描:2 天前 重新扫描
5 /100
daily-sleep
每日睡眠助手 — 晨间唤醒引导+睡眠质量回顾,晚间睡前放松程序+助眠技巧
Legitimate sleep assistant skill with no malicious behavior - uses filesystem only for user preferences, has proper path traversal protection, and outputs prompts to console.
技能名称daily-sleep
分析耗时26.8s
引擎pi
可以安装
No action needed. This skill is safe to use.

安全发现 3 项

严重性 安全发现 位置
低危
Missing permission declarations
SKILL.md does not explicitly declare filesystem READ/WRITE permissions for managing user preferences in data/users/
No allowed-tools or permissions section present
→ Add explicit permission declarations for filesystem:READ and filesystem:WRITE
SKILL.md:1
提示
Proper path traversal protection
All scripts implement safeUserPath() with directory boundary validation
if(!r.startsWith(path.resolve(USERS_DIR)+path.sep))
→ Good security practice - maintain this pattern
scripts/push-toggle.js:9
提示
Strict input validation
UserId sanitized with regex /^[a-zA-Z0-9_-]{1,128}$/ and time format validated
function sanitizeId(v){if(typeof v!=='string'||!/^[a-zA-Z0-9_-]{1,128}$/.test(v))
→ Good security practice - maintain strict validation
scripts/push-toggle.js:8
资源类型声明权限推断权限状态证据
文件系统 NONE READ ✓ 一致 push-toggle.js:7,20-21
文件系统 NONE WRITE ✓ 一致 push-toggle.js:28
网络访问 NONE NONE No network calls detected
命令执行 NONE NONE Node scripts execute locally, no shell commands executed
环境变量 NONE NONE No env access detected
数据库 NONE NONE No database access detected
1 项发现
🔗
中危 外部 URL 外部 URL
https://openclaw.ai
README.md:5

目录结构

7 文件 · 10.0 KB · 210 行
Markdown 2f · 110L JavaScript 3f · 85L JSON 2f · 15L
├─ 📁 scripts
│ ├─ 📜 evening-push.js JavaScript 18L · 1.5 KB
│ ├─ 📜 morning-push.js JavaScript 18L · 1.5 KB
│ └─ 📜 push-toggle.js JavaScript 49L · 3.5 KB
├─ 📋 _meta.json JSON 7L · 137 B
├─ 📋 package.json JSON 8L · 153 B
├─ 📝 README.md Markdown 44L · 1.6 KB
└─ 📝 SKILL.md Markdown 66L · 1.6 KB

依赖分析 2 项

包名版本来源已知漏洞备注
fs (built-in) N/A nodejs Node.js built-in module
path (built-in) N/A nodejs Node.js built-in module

安全亮点

✓ No network exfiltration or external IP connections
✓ No credential harvesting from environment variables
✓ No base64 encoded commands or obfuscated payloads
✓ No shell command injection vectors
✓ Proper path traversal protection with directory boundary checks
✓ Strict input validation for userId and time formats
✓ Whitelist approach for allowed channels (telegram, feishu, slack, discord)
✓ No remote script execution (curl|bash, wget|sh)
✓ Uses only Node.js built-in modules (fs, path) - no external dependencies with vulnerabilities
✓ User data stored locally in JSON format without sensitive information