nature-human-behaviour
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill is entirely informational, providing guidelines for academic manuscript framing, methodology, and formatting specific to Nature Human Behaviour.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze human behavioural science manuscripts (untrusted data).
- Ingestion points: User-supplied manuscripts and project descriptions analyzed in
SKILL.md. - Boundary markers: The instructions include a clear scope limiting its use to fit/venue selection and explicitly defer to official journal instructions.
- Capability inventory: The skill has no capabilities for file modification, command execution, or network exfiltration; it only produces text-based analysis.
- Sanitization: Not applicable as the skill performs no sensitive operations with the input.
- [EXTERNAL_DOWNLOADS]: The skill directs the agent to access external information from the official journal domain (nature.com) and suggests the use of established data repositories (OSF, Zenodo, GitHub, ICPSR). These are documented as standard practices for academic transparency and represent well-known, trusted services.
Audit Metadata