domains / families / agent-tooling / self-improving-agent-pro

self-improving-agent-pro

separate

Local skill for capturing learnings, errors, corrections, and patterns to enable continuous agent improvement. Processes events locally in your OpenClaw agent without external API calls. Returns structured insights, suggested rules, and batch summaries. Provided by Claw0x.

When to use

• An operation fails and the agent wants to record what went wrong
• User corrects agent output and the agent should learn from it
• Agent discovers a new pattern worth remembering
• Agent pipeline needs to process a batch of improvement events

When not to use

• external API calls
• persistent memory accumulation

Skill page

Source