ClawHub
Back to skill

Security audit

Stability AI

Security checks across malware telemetry and agentic risk

Overview

The skill appears to be a Stability AI connector, but its documentation frames it as read-only while describing an untagged action that generates audio and uploads a file.

Install only if you intend to let the agent generate Stability AI audio and upload the produced file. Treat calls as write/cost-incurring actions, require confirmation before generation, and review where generated files are stored or shared.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The manifest and description frame this skill as suitable for 'searching and reading data,' but the documented action performs generation and uploads a produced file. This mismatch can cause an invoking agent or user to treat the skill as read-only and run it without the confirmations normally required for state-changing or cost-incurring actions.

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The safety section states that untagged actions are reads and safe to run directly, yet the only listed untagged action generates audio and uploads it to transit storage. This creates a dangerous trust signal that could bypass user confirmation and lead to unintended content creation, storage, or billing-impacting operations.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The trigger text says to use this skill for ANY Stability AI request, which is overly broad and can cause an agent to invoke it even when a narrower, safer, or more appropriate path exists. Overbroad routing increases the chance of unintended execution of side-effecting actions, especially given the misleading read-only framing elsewhere in the file.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.