gemini-vision
Warn
Audited by Snyk on Feb 20, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md (usage examples like "Analyze from URL") and references/examples.md explicitly show downloading and analyzing images from arbitrary URLs (e.g., requests.get on a public image), so the agent ingests untrusted, user-controlled third-party content (images) that can be interpreted and used to drive follow-up prompts/actions.
Audit Metadata