Self-Improvement Skill

Enables any agent to learn and improve from user feedback using PRM-style evaluation.

What It Does

• Capture Feedback: Store user responses to agent outputs
• Evaluate: Extract score (+1 positive, -1 correction)
• Analyze: Find patterns in directive hints
• Suggest: Generate actionable improvements

Concept (OpenClaw-RL Inspired)

From OpenClaw-RL paper:

"Next-state signals encode both evaluative and directive information about the preceding action."

Two signal types:

1. Evaluative: Did the output work? (binary score)