claude-code-tips
Warn
Audited by Snyk on Feb 16, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.90). The skill explicitly performs WebSearch/WebFetch to ingest and summarize open web content — including community sources like GitHub Discussions, Stack Overflow, Reddit, Twitter/X, blogs, and other public websites — which are untrusted, user-generated sources the agent will read as part of its workflow.
Audit Metadata