ClawHub
Back to skill

Security audit

stable-audio-2.5

Security checks across malware telemetry and agentic risk

Overview

This skill is a straightforward Replicate-based music generation helper, with the main caveat that prompts and generated outputs go through a third-party service.

Install only if you are comfortable sending generation prompts and related inputs to Replicate and fetching generated MP3 outputs from external URLs. Do not include secrets, regulated data, private customer material, or confidential project details in prompts unless that third-party processing is approved, and keep the Replicate API token in environment variables rather than committed files.

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

Low
Confidence
90% confidence
Finding
The skill directs users to send prompts and generation parameters to Replicate using an API token, but it does not explicitly warn that user-supplied content will be transmitted to an external third-party service. This is a real privacy/transparency issue rather than code execution risk: users may unknowingly submit sensitive project details, prompts, or metadata off-platform.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The document instructs agents to send prompts and other user-provided inputs to Replicate and then download returned output URLs, but it does not warn that this transmits user content to a third-party service or that outputs may reside on external infrastructure. In a skill that handles creative prompts, narration text, or images/audio, this can cause unintended disclosure of sensitive user data if operators or agents assume the processing is local or first-party only.

External Transmission

Medium
Category
Data Exfiltration
Content
## p-image-ideogram (Pruna deployment)

Deployment: `prunaai/p-image-ideogram-preview`  
Endpoint: `POST https://api.replicate.com/v1/deployments/prunaai/p-image-ideogram-preview/predictions`  
Required input: `prompt`  
Optional: `mode` (`very low` · `low` · `medium` default · `high` · `very high`), `aspect_ratio`, `image_size`, `width`, `height` (when `aspect_ratio=custom`), `seed`, `output_format`, `output_quality`
Confidence
90% confidence
Finding
https://api.replicate.com/

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.