sdd4j-ears-tests

Pass

Audited by Gen Agent Trust Hub on Oct 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external requirement statements (EARS) to generate Java code and annotations. This process presents a surface for indirect prompt injection if the requirement text contains malicious instructions intended to influence code generation. The skill mitigates this risk by instructing the agent to use explicit TODO placeholders or failing stubs when responses cannot be deterministically mapped from the spec.
  • [DYNAMIC_EXECUTION]: The skill facilitates the creation of executable Java code, specifically requirement annotations and JUnit tests. The generation logic is structural and based on predefined templates, with the final execution occurring within the user's standard development environment rather than at skill load time or via unsafe runtime mechanisms.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 7, 2026, 01:15 PM