pattern-detector

Pattern detection scans session logs and memory to identify recurring command sequences, file access clusters, and similar prompts, surfacing potential skills and new agent ideas for automation.

Pattern detection for skill improvement is the process of identifying repeated usage behaviors across commands, files, prompts, and errors to queue pattern-driven skill enhancements and new automations.

You should run pattern detection during nightly optimization or evolve sessions to discover patterns across recent logs and memory, automatically queuing pattern-driven skill improvements for the optimizer.

Detecting usage patterns requires session logs and memory data sources, specifically capturing repeated command sequences, file access clusters, prompt similarities, and error occurrences to define detection logic.

Yes, you can surface new skill ideas by detecting repeated command sequences and similar prompts, logging the results, and notifying optimizers to queue pattern-driven skill improvements.

Pattern detection requires no external dependencies or components, operating directly on existing session logs and memory to define data sources, detection logic, output schema, and integration steps.