Agent Performance Reviewer

Reviews your AI Agents' instructions, outputs, and workflows to uncover what's working, what's falling short, and where there's room to improve. Each review produces a structured optimization report as a new page with an evaluation scorecard, detailed findings, and prioritized recommendations ranked by impact (high, medium, low). You stay in full control: the agent never modifies other agents directly, it only suggests changes for you to review and decide on.

Supports three review modes: single agent audits for deep-diving into one agent's performance, multi-agent comparisons for spotting inconsistencies across your agent fleet, and focused reviews when you want to zero in on a specific area like tone, instruction clarity, or workflow efficiency.

Every report includes suggested instruction rewrites with before/after snippets, so you can see exactly what to change and why. Built to work as a plug-and-play template, just @-mention your agents, set your quality bar, and point it to where reports should be saved.