research
Pass
Audited by Gen Agent Trust Hub on Feb 18, 2026
Risk Level: SAFE
Full Analysis
- Indirect Prompt Injection (SAFE): The research template in references/document-template.md lists external sources like web search and documentation as ingestion points. This represents a known attack surface for AI agents but is considered safe here as the skill provides methodology rather than implementing unsafe processing.
- Data Exposure & Exfiltration (SAFE): references/context-discovery.md includes a curl example using placeholder variables ($KEY, $ETL_URL). This is a standard instructional pattern for internal analytics and does not represent a concrete exfiltration threat.
- Command Execution (SAFE): scripts/validate.sh performs local file and content checks to verify skill structure. The script uses safe bash practices and does not execute untrusted external input.
Audit Metadata