NYC

wechat-article-writer

Warn

Audited by Snyk on Feb 15, 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). The skill's Step 1 ("搜索资料") explicitly instructs using a WebSearch tool to fetch and deeply summarize content from open public sources such as X/Twitter, Reddit and technical forums, so the agent will ingest untrusted, user-generated third‑party content as part of its workflow.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 15, 2026, 08:10 PM