ClawHub
Back to skill

Security audit

Cron Model Fix

Security checks across malware telemetry and agentic risk

Overview

This skill is a focused troubleshooting helper that changes an OpenClaw model allowlist and does not show hidden data collection or unrelated behavior.

Before running the fix, confirm the model name is one you intend to allow for agents and cron jobs. Prefer --dry-run first, keep the generated backup, validate ~/.openclaw/openclaw.json after edits, and restart the gateway only when ready for the policy change to take effect.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The manual editing section tells users to edit ~/.openclaw/openclaw.json directly and restart the gateway, but it does not first instruct them to back up the file or warn that malformed JSON or incorrect model entries can break agent behavior. In this context, the skill is specifically targeting a critical runtime configuration, so omission of backup/safety guidance can cause service disruption or require manual recovery.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.