ClawHub

Uncle Matt

Security checks across malware telemetry and agentic risk

Overview

This is a coherent security-wrapper skill that limits agents to approved Broker actions, with the main caveat that its optional tone pack and referenced external installer should be reviewed before use.

Before installing, inspect the referenced UNCLEMATTCLAWBOT repo and installer scripts because they configure OpenClaw and start a local Docker Broker. Keep the Broker loopback-only, avoid putting secrets in JSON config, and leave the voice pack disabled in professional or user-facing environments.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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 (3)

Natural-Language Policy Violations

Medium
Confidence
94% confidence
Finding
The skill description uses aggressive profanity in normal documentation rather than in a clearly gated, opt-in context. In agent ecosystems, this can surface directly to end users or operators, creating harassment, professionalism, and policy-compliance risks even if it does not directly compromise confidentiality or integrity.

Natural-Language Policy Violations

Medium
Confidence
97% confidence
Finding
A beginner guide title containing abusive/profane language is a true content-safety issue because it normalizes hostile wording in standard documentation that operators are expected to use. While not a direct technical exploit, it can cause trust, compliance, and workplace-appropriateness problems and increases the chance that similar language leaks into generated responses or support workflows.

Natural-Language Policy Violations

Medium
Confidence
89% confidence
Finding
The file hard-codes a strong persona and response style, including repeated profanity such as 'some stupid shit' and 'this is stupid,' without any user-facing opt-in beyond a feature flag. While this is not a direct code-execution or data-exfiltration flaw, it can cause policy-compliant refusals to be delivered in abusive or unprofessional language, creating reputational, trust, and safety issues in downstream deployments.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal