compose-page-blocks

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill mandates verbatim reproduction of all text nodes from the source HTML into generated block markup, so if the captured HTML (or related input files) contains API keys, tokens, cookies, or other secrets the LLM is required to output them unchanged, creating a high exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.85). The required runtime inputs include htmlPath pointing to a sanitized rendered HTML file from the source page; that HTML is outsider-authored content (public web page body text) and is read/parsed to generate post_content, so it can feed arbitrary free text into the agent’s LLM context.