last30days
Warn
Audited by Snyk on Feb 15, 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). This skill explicitly fetches and synthesizes user-generated content from public sources (Reddit and X) and general web pages — see SKILL.md ("Research ANY topic across Reddit, X, and the web") and scripts/last30days.py together with scripts/lib/openai_reddit.py, scripts/lib/xai_x.py and the websearch integration — and the agent is required to read and interpret those third‑party posts as part of its workflow, which can enable indirect prompt injection.
Audit Metadata