read-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 and parse content from public Fumadocs sites (via /llms.txt, raw .mdx pages, and the /api/search endpoint), meaning the agent will read and act on arbitrary third-party documentation that could contain untrusted or user-controlled instructions.

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

• Potentially malicious external URL detected (high risk: 1.00). This skill fetches remote documentation at runtime (e.g., https://example.com/llms.txt, https://example.com/docs/installation.mdx, and https://example.com/api/search?query=...) and injects/parses that raw MDX/JSON as model context to drive its answers, which constitutes remote content directly controlling agent prompts.