nd-emotion-regulation
Pass
Audited by Gen Agent Trust Hub on Aug 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security vulnerabilities or malicious patterns were identified in the provided files.
- [DATA_EXFILTRATION]: There are no network operations, URL references, or commands that attempt to access or export sensitive user data.
- [REMOTE_CODE_EXECUTION]: The skill is purely instructional and does not include any scripts, package dependencies, or mechanisms for executing remote code.
- [PROMPT_INJECTION]: The instructions do not contain markers for bypassing safety filters or overriding system instructions. Instead, they provide specific guardrails to prevent the agent from validating unverified interpretations or acting as a replacement for professional crisis support.
- [INDIRECT_PROMPT_INJECTION]: While the skill processes user input related to emotional states, it incorporates safety logic that requires separating observed facts from interpretations, which helps mitigate potential manipulation through emotional narratives.
Audit Metadata