research-lookup

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly calls third-party search APIs (Perplexity sonar-pro-search via OpenRouter in research_lookup.py _perplexity_lookup and the Parallel Chat API in _parallel_lookup) and returns/consumes public URLs, snippets and citation data (see SKILL.md and README) which the agent reads and uses to prioritize and generate research outputs, so untrusted web content can materially influence its actions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill makes runtime calls to OpenRouter (e.g., https://openrouter.ai/api/v1/chat/completions) and uses the API responses (notably the Gemini review "critique") to automatically modify and generate subsequent prompts (required via OPENROUTER_API_KEY), so external content directly influences agent prompts.