docs-seeker

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

• Third-party content exposure detected (medium risk: 0.30). Outsider-authored free text is ingested when the workflow reads externally fetched llms.txt and then WebFetch/Explorer agent URL contents or clones third-party GitHub repositories (via Repomix) selected from WebSearch results.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly WebFetches llms.txt files at runtime and then launches Explorer agents based on the llms.txt contents (e.g., https://docs.astro.build/llms.txt), so remote llms.txt content can directly control agent instructions.