content-repurposer

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

• Third-party content exposure detected (high risk: 1.00). The repurpose.sh workflow explicitly fetches arbitrary https:// URLs (using clawdbot web_fetch or curl) and then injects that fetched page text into the LLM prompts in the platform scripts (e.g., twitter-thread.sh, linkedin-post.sh), meaning untrusted third‑party web content is read and directly influences agent outputs.

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

• Potentially malicious external URL detected (high risk: 1.00). The main script accepts arbitrary https:// URLs (e.g., https://yourblog.com/post) and at runtime fetches their content via clawdbot web_fetch or curl (curl -sL "$SOURCE_INPUT" | html2text) and injects that fetched content directly into the LLM prompts, so external URLs can directly control agent instructions.