#![allow(missing_docs)]
pub const EXTRACT_EXPERIENCES_TEMPLATE: &str = r#"# Task: Extract Novel Learning Experiences
Analyze this conversation for novel learning experiences that would be surprising or valuable to remember.
## Conversation context
{{conversation_context}}
## Existing similar experiences
{{existing_experiences}}
## Instructions
Extract ONLY experiences that are:
1. Genuinely novel (not in existing experiences)
2. Actionable learnings about how things work
3. Corrections of previous mistakes or assumptions
4. Discoveries of new capabilities or patterns
5. Surprising outcomes that contradict expectations
Focus on technical knowledge, patterns, and cause-effect relationships that transfer to other contexts.
Avoid personal details, user-specific information, or routine interactions.
Respond with JSON array of experiences (max 3):
[{
"type": "DISCOVERY|CORRECTION|SUCCESS|LEARNING",
"learning": "What was learned (generic, transferable)",
"context": "What situation triggered this (anonymized)",
"confidence": 0.0-1.0,
"reasoning": "Why this is novel and valuable"
}]
Return empty array [] if no novel experiences found."#;