docs

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to "Fetch the documentation from the matching URL in the table below" and to "Answer the user's question using the fetched documentation," which means the agent will ingest and act on content from public third-party documentation URLs (e.g., https://mastra.ai/docs/llms.txt and other listed sites) that could contain untrusted/user-generated instructions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill fetches documentation at runtime from external URLs (e.g., https://mastra.ai/docs/llms.txt and the other listed llms*.txt URLs) and injects that fetched text into the agent's context to generate answers, so the remote content can directly control prompts/instructions.