openclaw-healthcheck 安全扫描报告 — 可信 | ClawSafe

5 /100

openclaw-healthcheck

Lightweight operational and security review for an OpenClaw deployment — checks gateway reachability, exposed ports, config patterns, browser session/relay surface, log errors, and runtime hygiene.

A legitimate OpenClaw operational health-check tool with no malicious behavior — all filesystem and shell access is directly declared and purpose-appropriate for the stated audit functionality.

技能名称openclaw-healthcheck

分析耗时27.2s

引擎pi

✓

可以安装

No action needed. The skill is safe to run as described.

安全发现 1 项

严重性	安全发现	位置
低危	Missing allowed-tools declaration 文档欺骗 SKILL.md declares a command-based workflow but does not include an allowed-tools mapping. The permission boundary is therefore implicit rather than explicit. `---` → Add an allowed-tools declaration to SKILL.md for transparency, e.g., 'allowed-tools: Bash→shell:WRITE, Read→filesystem:READ, fs.readFileSync→filesystem:READ'.	`SKILL.md:1`

资源类型	声明权限	推断权限	状态	证据
命令执行	`WRITE`	`WRITE`	✓ 一致	execSync('openclaw status'), execSync('lsof ...'), execSync('tail ...') — all de…
文件系统	`READ`	`READ`	✓ 一致	fs.readFileSync(configPath) and log tail — explicitly reading runtime state for …
环境变量	`NONE`	`READ`	✓ 一致	os.homedir() is called to resolve ~/.openclaw; environment is passed through to …

目录结构

6 文件 · 12.4 KB · 426 行

Markdown 3f · 270L JavaScript 2f · 122L JSON 1f · 34L

├─ ▾ 📁 references

│ └─ 📝 checklist.md Markdown 24L · 731 B

├─ ▾ 📁 scripts

│ └─ 📜 healthcheck.mjs JavaScript 80L · 3.5 KB

├─ ▾ 📁 tests

│ └─ 📜 test.mjs JavaScript 42L · 1.3 KB

├─ 📋 package.json JSON 34L · 832 B

├─ 📝 README.md Markdown 189L · 4.4 KB

└─ 📝 SKILL.md Markdown 57L · 1.7 KB

依赖分析 1 项

包名	版本	来源	已知漏洞	备注
`none (pure Node.js stdlib)`	`N/A`	stdlib only	否	healthcheck.mjs imports only child_process, fs, and os from the Node.js standard library

安全亮点

✓ No credential harvesting or environment variable enumeration for sensitive keys

✓ No network requests to remote endpoints — all checks are local

✓ No obfuscation (no base64, eval, or dynamic code construction)

✓ No download or remote script execution (no curl|bash, wget|sh, pip install from untrusted sources)

✓ No access to sensitive paths such as ~/.ssh, ~/.aws, or .env files

✓ No data exfiltration — evidence objects stay in-process and are output as JSON

✓ No persistence mechanisms (no cron, systemd units, or startup hooks injected)

✓ The test file properly isolates its execution by creating a temporary HOME and PATH

✓ Code behavior is fully consistent with the SKILL.md documentation