Reflection-Based Error Recovery

Apply reflection-based error recovery to handle LLM failures gracefully.

## Failed Operation
{{failed_operation}}

## Error Details
{{error_details}}

## Recovery Options
{{recovery_options}}

Use reflection to recover:

**Initial Reflection**
Analyze the failure:
- What went wrong?
- Was the error in input, processing, or output?
- Is this a transient or systematic issue?

**Diagnosis**
Reflect on the root cause:
- Was the prompt unclear?
- Was the input malformed?
- Was the model inappropriate for this task?

**Recovery Strategy Selection**
Choose the best recovery approach:
- Retry with modified prompt
- Fallback to alternative model
- Graceful degradation
- Request clarification

**Execute Recovery**
Apply selected strategy and verify success.

**Post-Recovery Reflection**
- Did recovery succeed?
- What can we learn for prevention?
- Should we update error handling?