Symptom Diagnosis Learning Skill: Conservative Refinement with Evidence-Based Addition

What problem does it solve?

This Skill helps agents refine symptom-to-diagnosis context without overfitting by adding only those discriminators that demonstrably improve validation performance.

Core Features & Use Cases

• Evidence-based discriminator addition: Adds new decision discriminators only when multiple validation gates strongly support benefit.
• Anti-overfitting gap preservation: Enforces a strict train-val gap constraint so improvements do not come from memorization.
• Strict error categorization and stopping rules: Categorizes errors using conservative criteria and stops when additions are unlikely to help.

Quick Start

Use the skill to load the existing symptom-diagnosis context, analyze train errors, and propose up to two new evidence-backed discriminators while preserving the train-val gap.