About
This skill automates the post-mortem analysis of failed agentic tasks within the AppWorld environment. By examining error logs, failed code snippets, and execution context, it identifies the root cause—such as API misuse, incorrect authentication sequences, or non-existent method calls—and transforms these failures into reusable, high-quality knowledge bullets. It ensures that future agent executions learn from past mistakes by providing concrete code examples and verified patterns for various integrated services like Spotify, Venmo, and Google Workspace.
