ClawHub
Back to skill

Security audit

Clawfy

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only skill that suggests ClawHub skills from browser or conversation context, with its proactive behavior disclosed and bounded.

Install only if you are comfortable with automatic skill suggestions during technical conversations. The skill may use recent conversation context and page metadata to form ClawHub search queries, so avoid enabling it where sensitive project names, private URLs, or confidential topics could appear in those queries.

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)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The skill description declares that it will proactively trigger during general technical discussion such as building, automating, deploying, integrating, or monitoring. This is overly broad for a passive assistant skill because it can activate without a clear user request, causing unsolicited processing of recent conversation context and external skill discovery in situations where the user may not expect it.

Vague Triggers

Medium
Confidence
97% confidence
Finding
Mode 2 instructs the agent to run `clawhub search` whenever the user discusses broad categories like building, automating, deploying, monitoring, or researching something technical. Because these conditions are ambiguous and common, the skill may perform external lookups and append recommendations during ordinary conversation, expanding data exposure and creating behavior that is difficult for users to predict or control.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal