sonarcloud

Warn

Audited by Snyk on Mar 13, 2026

Risk Level: MEDIUM
Full Analysis

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

  • Third-party content exposure detected (high risk: 0.70). The SKILL.md explicitly instructs the agent to query and proxy SonarCloud (third‑party) data via Membrane commands such as "membrane action list", "membrane action run", and "membrane request CONNECTION_ID /path/to/endpoint", meaning the agent will fetch and interpret user-generated SonarCloud/project content from the public service which could influence subsequent actions.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 13, 2026, 08:19 AM
Issues
1