java-pro
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): The instructions are standard role-definition constraints. There are no attempts to bypass safety filters or ignore system instructions.
- [Data Exposure & Exfiltration] (SAFE): No hardcoded secrets, API keys, or access to sensitive local file paths (e.g., ~/.ssh) were found.
- [External Downloads & RCE] (SAFE): The skill does not perform any network operations (curl, wget) or download external dependencies.
- [Command Execution] (SAFE): No shell commands or subprocess executions are present in the skill definition.
- [Indirect Prompt Injection] (LOW): The skill is designed to process Java code and architecture requirements. While it does not explicitly define boundary markers for untrusted input, the skill lacks any 'active' capabilities (tools, file writing, or network access) that could be exploited via malicious input data.
Audit Metadata