141-java-refactoring-with-modern-features

Pass

Audited by Gen Agent Trust Hub on Apr 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill mandates the use of shell commands such as ./mvnw compile, mvn compile, and mvn clean verify. These commands are used to ensure the project state is valid before and after refactoring, which involves executing the build lifecycle and any associated tests.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as categorized under Category 8. * Ingestion points: The skill processes external Java source code and Maven project files (pom.xml) which are treated as data to be refactored. * Boundary markers: There are no explicit delimiters or instructions provided to the agent to distinguish between the skill's instructions and potentially malicious instructions embedded in the user-provided code. * Capability inventory: The skill makes use of shell command execution capabilities to perform build and test operations. * Sanitization: The skill does not perform any validation or sanitization of the Maven configuration or the source code before invoking the compiler or test runner.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 27, 2026, 06:08 AM