agent-trace-analyzer

To debug AI agent decisions from session logs, analyze the traces to explain each decision point, identify root causes of failures like missing instructions or context rot, and receive actionable remediation suggestions.

Your AI agent fails correctly completing tasks due to root causes like missing instructions, context rot, wrong tool selection, or hallucinations, which are identified by analyzing agent session logs and decision points.

The best way to troubleshoot AI agent hallucinations and wrong tool selection is to analyze the session trace to pinpoint the exact decision point causing the failure, then apply targeted fixes to tool configurations.

To fix context rot in AI agent session traces, analyze the logs to classify the issue at the decision point, then apply targeted remediation planning by updating instructions in your CLAUDE.md file.

Yes, you can use session log analysis for remediation planning in CLAUDE.md, as the process identifies missing instructions and provides targeted fixes to update your tool configurations and prevent future failures.

Troubleshooting agent failures using session logs requires the analysis of complete decision points and traces, meaning limitations arise if session logs are incomplete or lack the context needed to identify the root cause.