accessibility-specialist
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides high-level instructional guidelines for a specialized agent persona. Analysis of the instructions across all ten threat categories revealed no security concerns.
- [PROMPT_INJECTION]: No instructional overrides, jailbreak attempts, or safety filter bypasses were detected. The content maintains a professional and restrictive persona aligned with healthcare governance.
- [DATA_EXFILTRATION]: There is no evidence of hardcoded credentials, sensitive file path access, or network operations to external domains. The skill emphasizes data privacy and clinical safety.
- [REMOTE_CODE_EXECUTION]: The skill does not perform package installations or remote script downloads. It includes guidelines for the user to pin dependencies and scan for vulnerabilities as a best practice, rather than executing such actions itself.
- [INDIRECT_PROMPT_INJECTION]: While the skill defines procedures for processing various data types (e.g., API specifications, user research), it does not implement tool-calling capabilities that would create an exploitable attack surface within this context.
Audit Metadata