Verifies AI agent work end to end: real artifacts and outcomes checked, not self-reported success.
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.
Verifies AI agent work end to end: real artifacts and outcomes checked, not self-reported success.