research-synthesis
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified in the skill instructions or workflow. The skill is designed to process TTRPG-related research material and generate design suggestions.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external source material (notes, transcripts, documents). While this represents a potential surface for indirect prompt injection, the skill lacks access to high-risk tools such as network operations or arbitrary code execution, and its internal logic emphasizes separating evidence from interpretation, which serves as a natural boundary marker.
- [COMMAND_EXECUTION]: While the skill instructions mention writing files to a repository, there are no patterns suggesting the execution of unauthorized or dangerous shell commands. The operations described are consistent with standard agentic file management tasks.
- [DATA_EXFILTRATION]: There is no evidence of hardcoded credentials, sensitive file path access, or network calls to unauthorized domains.
Audit Metadata