reduck

Warn

Audited by Snyk on May 12, 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). SKILL.md explicitly instructs the agent to call the Reduck CLI to scrape and fetch arbitrary public websites and social platforms (e.g., "scrape a site, extract structured records, search LinkedIn/Reddit/Google" and the "Go to " reduck run -q example), then read and act on the returned records as part of its workflow, which exposes it to untrusted third‑party content that could carry injected instructions.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
May 12, 2026, 09:36 AM
Issues
1
Security Audit — snyk — reduck