code-review-architect
Pass
Audited by Gen Agent Trust Hub on Apr 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a set of natural language instructions for a code review assistant. It does not include any scripts, executable commands, or configuration that could lead to unauthorized actions.
- [PROMPT_INJECTION]: No evidence of instructions attempting to bypass safety filters or override agent behavior was found. The instructions are task-oriented and professional.
- [DATA_EXFILTRATION]: The skill does not perform any network operations or access sensitive local files. It identifies 'Data Exposure' as a risk to check for during code reviews, which is a positive safety feature.
- [REMOTE_CODE_EXECUTION]: There are no patterns involving remote script downloads (e.g., curl|bash) or package installations.
- [INDIRECT_PROMPT_INJECTION]: While the skill is designed to process untrusted data (code diffs from users), it does not possess exploitable capabilities such as file writing or network requests that could be abused by malicious payloads in the input code.
Audit Metadata