deep-research-pro

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md workflow explicitly instructs the agent to run DuckDuckGo searches (via /home/.../ddg) and to curl/fetch full web/news URLs for "deep reading" (Steps 3–4), meaning it ingests untrusted public web content that the agent reads and uses to drive synthesis and follow-up actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill clearly fetches arbitrary external webpages at runtime (e.g., the SKILL's curl pipeline curl -sL "" | python3 -c "..." and the CLI --fetch example "https://example.com/article"), and that fetched content is read into the agent's context to drive synthesis, so remote URLs can directly control prompts.