Agent Error Recovery Rate Calculator

Agent Error Recovery Rate Calculator MCP Connector for Claude

A+

Analyzes execution logs to calculate tool call success ratios and agent recovery rates.

3 tools Official Updated Oct 1, 2026 Official Vinkius Partner

This MCP server provides specialized tools to analyze the resilience of AI agents by parsing execution logs. It identifies specific error signatures and computes critical performance metrics, such as the success-to-failure ratio and the recovery rate. Use parse_execution_logs to structure raw logs, calculate_error_distribution to identify error patterns, and calculate_resilience_metrics to determine how effectively an agent recovers from failures.

logserror-analysisresiliencemetricsdebugging

3 tools expose this connector's capabilities to your AI agent.

calculate_resilience_metrics

Computes the high-level performance ratios regarding reliability and recovery

parse_execution_logs

Extracts a structured sequence of tool call outcomes and error types from raw text logs

calculate_error_distribution

Provides a breakdown of how many times specific errors occurred

See how to talk to your AI agent using Agent Error Recovery Rate Calculator.

Can you analyze these logs and tell me the recovery rate?

The recovery rate for the provided logs is 75%, meaning 3 out of 4 failures were followed by a successful tool call.

What are the most common errors in my execution history?

The most frequent error identified is 'Timeout', which occurred 5 times in the analyzed logs.

Parse these logs for me.

I have parsed the logs. I found 10 total events: 8 successes and 2 failures.

You can calculate the success-to-failure ratio and the recovery rate percentage using `calculate_resilience_metrics` after parsing your logs.

Related Connectors