ClawHub
Back to skill

Security audit

sogni-worlds

Security checks for vulnerabilities and agentic risk

Overview

The skill coherently builds interactive worlds using AI services, with user controls for spending, review and public sharing.

Installation downloads and runs the external kit. Before rendering personal photos, check the services' privacy policies and choose images you are comfortable sending for processing. Review costs and approve public hosting or gallery sharing separately.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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 (1)

The photo-world rendering workflow lacks a disclosure about third-party processing of user imagery.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The skill supports generating worlds from personal photos using external AI services but does not explicitly require disclosure and informed consent before imagery is sent for processing. Photos may reveal faces, private locations, or other sensitive information; approval of rendering costs or public gallery submission does not establish consent to third-party image processing. The delegated AGENTS.md workflow might provide safeguards, but its contents are unavailable, so that protection cannot be verified.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.