legacy-modernizer

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWPROMPT_INJECTIONNO_CODE
Full Analysis
  • PROMPT_INJECTION (LOW): The skill possesses a vulnerability surface for indirect prompt injection due to its core function of processing untrusted external codebases.\n
  • Ingestion points: Legacy codebases, framework files, and database schemas identified in the Focus Areas.\n
  • Boundary markers: Absent; there are no instructions to use delimiters or to disregard embedded natural language instructions within the processed code.\n
  • Capability inventory: The skill is capable of reasoning over and generating refactored code, migration plans, and test suites.\n
  • Sanitization: Absent; no methods for validating or escaping external content are specified.\n- NO_CODE (INFO): The skill is purely instructional and contains no executable code, remote dependencies, or binary components, which significantly limits the potential for direct system compromise.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 04:57 AM