Triage large log files for errors
A bounded Markdown summary of a large log file, including error levels, the first failure, grouped messages, stack traces, and the time span, enabling quick incident assessment.
How it works
The agent uses the 'Triage' skill from the Agentic Keychain MCP server. It streams a large log file to the server, which processes it to print a bounded Markdown summary. The agent should check the server's response for the summary details, specifically looking for error levels, the first failure, recurring messages, stack traces, and the time span to understand the log's content without reading the entire file.
Prompt
Paste this into an agent that has the servers below connected. Adjust the names and dates to your own.
Please triage the following large log file and provide a bounded Markdown summary including error levels, the first failure, grouped messages, stack traces, and the time span:
[Insert Log Content Here]MCP servers it needs
Source
The vendor page this playbook is drawn from, and the passage that supports it.
Triage Streams a large log once and prints a bounded Markdown summary - levels, first error, grouped messages, stack traces, time span.Agentic Keychain—Agent Skill ROI Registry & Benchmarking