web-search

Warn

Audited by Snyk on Nov 18, 2026

Risk Level: MEDIUM
Full Analysis

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

  • Third-party content exposure detected (low risk: 0.10). The required workflow runs Tavily and Exa search/extract apps that fetch arbitrary public web pages, which the agent selects by query or URL, and the returned page text is fed to an LLM, so this is low risk because outsiders can publish poisoned pages but cannot force delivery into the workflow.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

  • Potentially malicious external URL detected (medium risk: 0.30). The decisive boundary is the setup command on line 7, npx skills add belt-sh/cli, which is an instructed remote-repository fetch that installs skill material; the publisher belt-sh is not clearly established as the inference.sh first-party organization, so it is treated as common/reputable unofficial and rated medium, while the belt app run calls are control-plane and their outputs are data.

Issues (2)

W011
MEDIUM

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

W012
MEDIUM

Unverifiable external dependency detected (runtime URL that controls agent).

Audit Metadata
Risk Level
MEDIUM
Analyzed
Nov 18, 2026, 08:44 PM
Issues
2