agent-onboarding

Warn

Audited by Snyk on Apr 26, 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). This skill instructs agents to fetch and ingest arbitrary public web content as part of its required workflow (e.g., via "bdata scrape"/"bdata search", the Web Unlocker REST POST to https://api.brightdata.com/request, site pipelines for platforms like Reddit/LinkedIn/Instagram, and the MCP URL https://mcp.brightdata.com), meaning untrusted third-party/user-generated content can directly influence agent tool use and decisions.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 26, 2026, 11:11 AM
Issues
1