markitdown

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and converts arbitrary public URLs (including YouTube, Wikipedia, RSS feeds, and "general web URLs") into Markdown and then uses that content to answer the user's questions, so it ingests untrusted third‑party web/user-generated content that could carry indirect prompt-injection instructions (see "When to Use" and the Quick Reference/Examples in SKILL.md).

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly recommends running a runtime install command that fetches and executes remote code — "curl -LsSf https://astral.sh/uv/install.sh | sh" — to install the required uvx dependency, which meets the criteria for a high-risk external URL.