axiom-audit

What problem does it solve?

Manually sifting through Axiom logs to find, prioritize, and understand errors and warnings is a tedious and reactive process. This skill automates log auditing to proactively identify and research issues, reducing incident response time and improving system reliability.

Core Features & Use Cases

• Automated Error & Warning Querying: Connects to Axiom to query logs for errors and warnings within specified time ranges using APL.
• Prioritized Issue Reporting: Scores issues by severity and frequency (P0-P4), generating a prioritized list for immediate action.
• Root Cause Research & Log Smell Detection: Researches probable causes in the codebase and flags common log anti-patterns like excessive logging or sensitive data exposure.
• Use Case: Run this skill daily to automatically audit production logs, identify any new or escalating critical errors, and generate a prioritized report for the engineering team, enabling proactive issue resolution.

Quick Start

Install axiom-mcp and mcptools (if not already installed)

go install github.com/axiomhq/axiom-mcp@latest brew install mcp # macOS

Set Axiom credentials

export AXIOM_TOKEN="xaat-your-token"

List datasets

mcp call listDatasets --params '{"arguments":{}}' ~/go/bin/axiom-mcp

Query errors for the last 24 hours

mcp call queryApl --params '{"arguments":{"dataset":"logs","apl":"['''now-24h''':now] | where level == "error" | summarize count() by message"}}' ~/go/bin/axiom-mcp