astro

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

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow fetches public Astro documentation text files (e.g., https://docs.astro.build/llms-small.txt and other _llms-txt/*.txt) via a research sub-agent, and that fetched free-form prose is then ingested into the agent’s LLM context—an outsider source (public web content).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs a spawned research sub-agent to fetch external documentation at runtime (e.g., https://docs.astro.build/llms.txt and the referenced https://docs.astro.build/llms-small.txt, https://docs.astro.build/llms-full.txt and _llms-txt/* files), which are then injected into the agent's prompt/context to drive its responses, so these URLs are required runtime dependencies that directly control agent instructions.