content-radar

Warn

Audited by Snyk on Jun 25, 2026

Risk Level: MEDIUM
Full Analysis

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

  • Third-party content exposure detected (high risk: 0.85). Outsider free text can enter the agent’s LLM context via the runtime web-search path: when last30days is not available, content-radar.sh emits ##SEARCH_DIRECTIVE## blocks for the agent to perform live searches on Reddit/X, and the agent then fills the “What’s Being DISCUSSED (Reddit + X)” section using real thread titles/snippets/quotes from outsider-authored posts.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 25, 2026, 09:45 PM
Issues
1
Security Audit — snyk — content-radar