Check whether an AI answer is grounded in its context — deterministic, no LLM judge.
Verification confirms publisher identity (repo ownership), not code safety. The security scan covers known CVEs and suspicious install scripts — it cannot prove the absence of malicious code.
Check whether an AI answer is grounded in its context — deterministic, no LLM judge.