agentic-ai-research-agent
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill retrieves its implementation code from the official GitHub repository of DeepLearning.AI, which is a well-known and reputable educational organization.
- [CREDENTIALS_UNSAFE]: The skill adheres to security best practices by using environment variables for managing sensitive API keys and recommending the use of .env files. It includes default credentials only for a local development database instance within a Docker container, which is standard practice for local-first development setups.
- [PROMPT_INJECTION]: The skill features an indirect prompt injection surface as it ingests untrusted user data via the
/generate_reportendpoint inmain.pywithout explicit boundary markers or sanitization logic. However, the risk is mitigated because the agents' capabilities are restricted to executing search queries against public academic and web APIs (Tavily, arXiv, Wikipedia) and recording task state in a local database.
Audit Metadata