Reflect on AppWorld Failure

Analyze failed AppWorld tasks to extract specific, actionable learnings that can be added to the playbook.

Purpose

When an AppWorld task fails, the Reflector calls this Skill with error details and failed code. You analyze the failure semantically and generate a high-quality bullet with:

1. Specific title describing the pattern
2. Detailed content with working code examples
3. Relevant tags for retrieval
4. Appropriate confidence level

Input Format

The input will be a text description with sections:

# Task
<task instruction>