ClawHub
Back to skill

Security audit

Bring Your Own Model

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward Runware model-import guide, but users should protect any private model-weight URL they provide.

Before using this skill, host model weights with the narrowest practical access: prefer short-lived signed URLs, avoid permanent public buckets, do not embed secrets in shareable URLs, expire or delete the hosted file after import, and keep Runware's private setting enabled unless publication is intentional.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

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 (1)

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill explicitly tells users to provide a publicly fetchable or signed URL for their model weights, but does not warn that this may expose proprietary, private, or licensed model files to unintended parties if the URL is misconfigured, logged, or shared. In this context, users are uploading custom model artifacts that are often sensitive IP, so omission of privacy guidance creates a realistic risk of accidental disclosure rather than a purely theoretical issue.

Static analysis

No suspicious patterns detected.