article-to-wechat-cover

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and ingests arbitrary Feishu/Lark documents (via fetch_lark_doc / lark-cli and load_lark_doc_article) and uses that user-provided document content in analyze_article_theme to build OpenRouter prompts and the execution_spec that directly control image-generation and downstream actions (see SKILL.md and scripts/article_to_wechat_cover.py), so untrusted third‑party text can materially influence tool behavior.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill makes runtime POST requests to the OpenRouter API at https://openrouter.ai/api/v1/chat/completions (via call_openrouter) and the returned model JSON is parsed and used directly to build the image-generation prompt/spec (and OPENROUTER_API_KEY is required), so this external URL at runtime directly controls prompts.