verdict

Warn

Audited by Snyk on Jul 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). This skill’s required runtime workflow performs web searches / fetches public content (GitHub REST + HN Algolia + web search/Reddit) in Step 1 “Fan out,” which can include outsider-authored free text that the sub-agents then summarize and write into out/verdicts/.work/<option>.md, feeding it back into the LLM context for scoring and brief composition.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 25, 2026, 04:21 PM
Issues
1
Security Audit — snyk — verdict