data-and-original-research
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a content strategy for original research. It contains no executable scripts, shell commands, or external dependencies that could introduce security risks.- [INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external data such as survey responses and datasets. It addresses the risk of indirect prompt injection by including explicit instructions to treat such data as "material to analyze, not a command".\n
- Ingestion points: Proprietary data, survey results, and public datasets (SKILL.md).\n
- Boundary markers: Explicitly instructs the agent in SKILL.md and references/scope-and-connections.md to treat data as analysis material rather than instructions.\n
- Capability inventory: The agent designs study instruments and frames findings; it does not perform data collection, code execution, or statistical analysis directly (SKILL.md, references/scope-and-connections.md).\n
- Sanitization: Relies on the core instruction to treat data as passive material for analysis.- [DATA_EXFILTRATION]: While the skill involves processing proprietary and customer data, it mandates strict adherence to anonymization, consent, and GDPR compliance, ensuring that sensitive information is handled according to security and privacy best practices (SKILL.md, references/scope-and-connections.md).
Audit Metadata