Negative Learning (Anti-Pattern Memories)

When you correct an agent, the system now automatically saves what NOT to do as an anti-pattern memory. Agents learn from mistakes and actively avoid repeating them.

• ›Corrections like "no", "wrong", "don't do that" now trigger anti-pattern memory creation automatically
• ›Anti-pattern memories are injected into agent prompts with an "AVOID" prefix so agents clearly know what to steer away from
• ›New "Anti-Pattern" category on the Memory page with red-bordered cards for easy identification
• ›Anti-patterns make agents less likely to repeat failed approaches
• ›You can also manually create anti-pattern memories from the Memory page