Critiqor observes AI agent runtime across Codex, Claude Code, OpenClaw, and custom CLI workflows, then turns runtime evidence into diagnoses your team can trust and act on.