content-repurposer

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

• Third-party content exposure detected (medium risk: 0.30). scripts/repurpose.sh accepts a user-supplied <source_file_or_url> argument (including arbitrary URLs), fetches/extracts its text into CONTENT, then passes that free text via stdin to the platform scripts (e.g., scripts/twitter-thread.sh) for LLM transformation.

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

• Potentially malicious external URL detected (high risk: 0.80). scripts/repurpose.sh will fetch arbitrary HTTP(S) source URLs at runtime (e.g., the example "https://yourblog.com/post") via curl or clawdbot web_fetch and the fetched content is injected into the LLM prompts in the platform scripts (e.g., twitter-thread.sh), allowing remote content to directly control model input.