error-memory

To stop repeating CLI errors, an error memory system automatically consults a learned-errors knowledge base after two consecutive failures, fetching and applying previously discovered solutions to prevent wasted time and token overhead.

Automatic error lookup triggers after two consecutive command failures, querying a structured knowledge base indexed by keywords and error categories to instantly retrieve a known resolution and apply it to the current workflow.

Yes, you can manually query the structured error memory to retrieve learned errors and apply saved resolutions across various environments, including Windows, Supabase, and CLI task workflows.

Learned errors are stored safely under .claude/memory/errors with an INDEX.md file, organizing entries into predefined error classes and applying keyword-based indexing for fast and structured lookups.

The best way to manage persistent errors in automated workflows is using a persistent error memory that categorizes failures into predefined classes and safely persists new entries with their solutions to an indexed knowledge base.

No external dependencies are required, as the error management system operates independently by safely persisting new error entries and an INDEX.md directly within the local .claude/memory/errors directory structure.