google-deep-research

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill's examples and workflow require embedding API keys directly into request headers and assigning them into code variables (and even show key-like strings), which forces the agent/LLM to handle and potentially output secret values verbatim.

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 sending queries to Google's Deep Research API which "runs comprehensive research queries server-side, searching multiple web sources" and returns synthesized reports with source citations/URLs (e.g., "Search ESG reports, press releases, LinkedIn..." in Query Design), so the agent consumes untrusted public web content that can materially influence its findings and actions.