ClawHub

test0413-6348

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed local OpenClaw skill audit tool; its broad file reading and local snapshots fit that purpose, but users should review scheduling and stored diff contents.

Install only if you want a local auditor that scans and copies your OpenClaw skills directory. Keep secrets out of skill files, confirm the notification channel and timezone before enabling cron, and do not grant wallet or OAuth credentials for this skill because the artifacts do not need them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Natural-Language Policy Violations

Medium
Confidence
84% confidence
Finding
The example cron command hard-codes `--tz "Asia/Shanghai"`, which can cause jobs to run at unexpected times if users copy it verbatim. While not a direct code-execution issue, forcing a timezone without user opt-in can mislead users and create operational or notification-routing problems.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The context profile system uses very broad, generic keywords such as 'run', 'file', 'text', 'monitor', and 'agent' to decide when to ignore or downgrade security rules. This can cause misclassification of unrelated skills and suppress detection of genuinely dangerous behavior, creating a policy-bypass condition in the scanner rather than directly executing malicious code.

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal