coding-standard-java
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- Prompt Injection (SAFE): No instructions found that attempt to bypass safety filters or override system behavior.
- Data Exposure & Exfiltration (SAFE): No hardcoded credentials, sensitive file paths, or network communication identified.
- Unverifiable Dependencies & Remote Code Execution (SAFE): The skill does not perform any package installations or download remote scripts.
- Indirect Prompt Injection (LOW): The skill processes external Java code (Ingestion Point: SKILL.md body instructions). Boundary markers are absent. Capability inventory reveals no subprocess calls, exec/eval calls, file-write operations, or network access across the skill definition. Sanitization is absent.
- Dynamic Execution (SAFE): There is no runtime code generation or execution of dynamic content.
Audit Metadata