blog-feedback

Warn

Audited by Snyk on Mar 10, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 1.00). SKILL.md explicitly says webpage URLs are fetched via "defuddle skill 或 WebFetch" and the agent must read the retrieved article section-by-section with the Read tool, so arbitrary public web content (untrusted user-generated pages) is ingested and interpreted as part of the workflow and could contain instructions that alter the agent's behavior.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 10, 2026, 04:28 PM