named-persona-adversarial-review
Pass
Audited by Gen Agent Trust Hub on Jul 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary function is to provide structured feedback on code using established engineering principles. No malicious instructions, such as prompt injections or bypass attempts, were identified.
- [SAFE]: No evidence of data exfiltration, credential exposure, or unauthorized network operations was found. The skill operates entirely within the provided context and documented references.
- [SAFE]: The skill does not perform any external downloads, remote code execution, or package installations. All dependencies (like the persona principles) are stored locally within the skill's file structure.
- [SAFE]: There are no indicators of obfuscation, persistence mechanisms, or privilege escalation attempts. The logic is transparent and focused on the stated goal of code analysis.
- [SAFE]: While the skill is designed to process external data (source code from PRs), it includes built-in mitigations like 'Attribution discipline' and an 'Integrity Check' to ensure findings are grounded in documented principles and technical merit, reducing the risk of indirect prompt injection.
Audit Metadata