mslm-cloud

Warn

Audited by Snyk on Apr 23, 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.md explicitly shows runtime flows that fetch and proxy data from Mslm Cloud (e.g., "File Content" in the overview and the "Proxy requests" section with the command membrane request CONNECTION_ID /path/to/endpoint), meaning the agent will ingest user-generated/untrusted third-party content from external cloud accounts that could materially influence its actions.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 23, 2026, 02:01 PM
Issues
1