wa:retrospective

To analyze AI session logs for root causes, classify observations into tool-pattern, workflow-tactic, error-pattern, and context-issue categories. Retrospective analysis extracts correction patterns from audit logs and error archives to identify recurring infrastructure failures.

Retrospective session analysis is the process of classifying session observations and extracting correction patterns from audit logs. You need it for lifecycle troubleshooting workflows requiring learning capture, deduplication, and pattern detection across tools and execution stages.

Detect repeated tool errors and oscillation patterns by running a deep forensic mode analysis on full audit logs. This deep mode identifies repeated file reads, tool oscillation, error clusters, and calculates tool frequency statistics for comprehensive troubleshooting.

Yes, you can automatically record learning captures and update error catalogs after troubleshooting. The process records findings into structured memory topic files, prunes the memory hub, and updates the error catalog via correction detection and deduplication.

Yes, retrospective analysis supports conditional remediation with user-approved artifact changes. It applies to lifecycle troubleshooting workflows that propose remediation targets and conditionally apply user-approved artifact changes to fix recurring active error patterns.

The best way to prevent recurring infrastructure failures using audit history is to perform retrospective analysis, record learnings into structured memory files, and propose remediation targets. This reduces recurring workflow failures by analyzing session errors and audit history.