ClawHub
Back to skill

Security audit

Belief Assessor

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent fact-checking helper, but users should avoid feeding it confidential claims or evidence unless external search and LLM processing are acceptable.

Install is reasonable for ordinary fact-checking. Before using it with private, regulated, proprietary, or personally sensitive material, redact inputs or run with approved/local search and LLM tooling, because the workflow may send claims and evidence outside the local environment.

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 (2)

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill mandates online searching as a default first step but does not warn that user claims, evidence, or contextual details may be sent to third-party search providers. In practice, this can cause unintended disclosure of sensitive, proprietary, regulated, or personal data and may violate privacy, compliance, or data-handling requirements.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill sends the full claim and up to 1500 characters of evidence into a caller-supplied LLM callback, which may be backed by a remote provider. If claims or evidence contain secrets, personal data, proprietary content, or sensitive investigation notes, this creates a real data-exposure path with no built-in consent, redaction, or locality controls. In this skill context, the risk is somewhat elevated because the purpose of the tool is to assess arbitrary evidence, which often includes sensitive text gathered from other systems or users.

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.