rtl-equivalence-checker

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by processing external RTL source files and incorporating their contents into analysis reports interpreted by the agent.
  • Ingestion points: User-provided Verilog files are read and parsed by 'scripts/rtl_parser.py'.
  • Boundary markers: The report formatting in 'scripts/check_equivalence.py' uses visual separators (e.g., dashes and headers) but does not include explicit instructions to the LLM to ignore embedded commands within the parsed RTL code.
  • Capability inventory: The skill identifies logic differences and generates counterexamples through local computations. It has the capability to write results to a user-specified file path, but no network operations or subprocess executions were identified.
  • Sanitization: The skill does not perform any escaping or sanitization of signal names, module names, or logic snippets extracted from the source files before displaying them in the output.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 10:21 PM