ClawHub

Self Review

Security checks across malware telemetry and agentic risk

Overview

This is a simple local output-quality checker with no evidence of data theft, persistence, hidden installs, or destructive behavior.

Safe to install as a lightweight local formatting and clarity check. Keep it advisory or set clear limits if you wire exit code 1 to regenerate responses, especially for safety-critical, factual, legal, medical, financial, or multilingual output.

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)

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The README explicitly states that outputs needing improvement will be automatically regenerated, which means user-facing content may be altered by a hidden review step. Without clear disclosure, constraints on what may be changed, and auditability, this can cause silent modification of agent responses, including safety caveats or factual details, before they reach users.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The integration guidance recommends placing this skill directly in the path of agent responses before they are sent to users, but it does not warn that the skill can modify outbound content. In that context, the gate becomes a silent post-processor that may distort meaning, remove disclaimers, or create responses the agent did not originally produce.

Natural-Language Policy Violations

Medium
Confidence
91% confidence
Finding
The heuristic explicitly checks for Chinese- and English-specific action words, which biases approval toward outputs written in those languages and can incorrectly fail valid responses in other languages. In a self-review gate, this can systematically suppress or force reshaping multilingual output, creating unfair quality gating and reliability issues for non-target languages.

VirusTotal

53/53 vendors flagged this skill as clean.

View on VirusTotal