Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

SysGuard

v2.1.1

实时监控OpenClaw系统CPU、内存、磁盘及网关状态,支持健康检查、诊断报告、趋势图和缓存清理。

0· 93·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
Name/description (monitor CPU/memory/disk/Gateway, health checks, diagnostics, trends, cache cleaning) align with the included shell scripts: health_check, diagnostics, trend, clean, monitor, notifier and UI formatting. The use of OpenClaw-specific commands (openclaw gateway probe/restart) is consistent with an OpenClaw system guardian.
!
Instruction Scope
Runtime instructions and scripts generally stay within monitoring and cleanup scope, but diagnostics/utility functions probe a hardcoded external endpoint (api.minimaxi.com) via curl/ping to measure API latency. This external network contact is not documented in SKILL.md and could leak reachability/latency information or be used as a beacon. The SKILL.md asserts 'zero-dependency' yet scripts rely on common utilities (curl, ping, pgrep, journalctl, bc, optionally jq) — jq is mentioned in code but not documented in the top-level description as an optional dependency.
Install Mechanism
No download/install spec is present (files included in the package). There are no external archives or installers pulled during install, so installation risk is low from a supply chain perspective. ClawHub one-click or git clone are the documented install options.
Credentials
The package declares no required env vars or credentials. It does, however, read and persist configuration (config/sysguard.conf) and supports FEISHU_WEBHOOK_URL / WECOM_WEBHOOK_URL for notifications — those webhook URLs would receive formatted system status and diagnostic messages. Defaults reference filesystem paths (/root/.openclaw, /mnt/vdc, /tmp for config when not run through the skill wrapper) which may be surprising; user should confirm/defaults are appropriate. No unrelated cloud credentials or secrets are requested.
Persistence & Privilege
The skill does not request permanent platform privileges (always:false). It can run a background monitor loop (nohup) if the user starts it, but it does not auto-enable itself across the system or modify other skills. It writes logs and history data under its data/logs and data/history directories (and can operate on /tmp/openclaw and system journal when cleaning).
What to consider before installing
What to check before installing: - The code appears to implement the advertised monitoring features, but review and (if needed) remove or replace the hardcoded external host (api.minimaxi.com) used for the 'API 响应超时' check — if you don't trust that host, it should be changed or removed. - Configure notification webhooks (FEISHU_WEBHOOK_URL / WECOM_WEBHOOK_URL) carefully — these receive system status and diagnostics. Only set them to endpoints you control. - Verify default paths (OPENCLAW_ROOT, DATA_DISK, BACKUP_ROOT, CONFIG_FILE) are appropriate for your system; the defaults point to /root/.openclaw and /mnt/vdc which might not exist or be sensitive. - The scripts call utilities such as curl, ping, pgrep, journalctl, bc and optionally jq; ensure those are available and acceptable on your host. 'jq' is required for trend charts if you want that feature. - The cleaner runs journalctl --vacuum-time and deletes files in /tmp/openclaw and historical JSONs; review the clean.sh logic and retention settings (HISTORY_RETENTION_DAYS) to avoid unintended deletions. - Run the scripts in a staging environment first and inspect what data is written and what network calls are made (especially outbound requests to webhooks and api.minimaxi.com). If you need, ask the author to document the api.minimaxi.com purpose or replace it with a configurable probe target. Overall: functionally coherent, but the undocumented external probe and a few imprecise claims ("zero-dependency") are reasons to inspect and possibly adjust configuration before production use.

Like a lobster shell, security has layers — review code before you run it.

latestvk97c7vd5t3ayrc0kj2h93gs2zs83gfyk

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments