prompt-builder

Warn

Audited by Snyk on Mar 21, 2026

Risk Level: MEDIUM
Full Analysis

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

  • Third-party content exposure detected (high risk: 0.80). The skill's reference guidance (references/prompt-builder.md, under "Tool Usage") explicitly instructs the agent to use $browser.search and cites public sites (e.g., reddit.com and other web links), meaning the agent will fetch and interpret open, user-generated web content that can materially influence prompt creation and tool use.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 21, 2026, 07:18 AM
Issues
1
Security Audit — snyk — prompt-builder