122-java-type-design

Pass

Audited by Gen Agent Trust Hub on Apr 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes shell commands (mvn compile, ./mvnw compile, mvn clean verify) to validate Java projects. While these are standard developer tools, the execution of the project-local Maven wrapper (./mvnw) involves running a script within the target project directory.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it analyzes and refactors user-supplied Java code.
  • Ingestion points: Java source code files provided by the user.
  • Boundary markers: Absent.
  • Capability inventory: Shell command execution via Maven tools as defined in SKILL.md and references/122-java-type-design.md.
  • Sanitization: The skill does not perform sanitization or isolation of the code being processed.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 23, 2026, 06:06 PM