System Reliability Engineering (SRE) skill

This page explains how to use the SRE skill to give AI coding agents expert investigation capabilities for incidents and debugging.

The SRE skill turns AI agents into expert SRE investigators:

Investigation methodology: Hypothesis-driven debugging with systematic triage using Golden Signals and USE/RED methods
APL query patterns: Reference materials for constructing effective Axiom queries
Memory system: Persistent storage that learns from debugging sessions, capturing what works and what doesn't
API helpers: Scripts for querying Axiom directly from the agent

Prerequisites

Create an Axiom account.
Create a dataset in Axiom where you send your data.
Create an API token in Axiom with permissions to ingest data to the dataset you have created.

Install Axiom Skills

Install all Axiom skills at once for Claude Code, Cursor, and other Claude-compatible agents:

shell

npx skills add axiomhq/skills

Configure Axiom credentials

All Axiom Skills share the same credential configuration. Create a configuration file at ~/.axiom.toml:

TOML

[deployments.dev]
url = "https://api.axiom.co"
token = "API_TOKEN"
org_id = "ORGANIZATION_ID"
edge_url = "AXIOM_DOMAIN"

Use SRE skill

The SRE skill activates automatically when you ask your AI agent about:

Incident response and debugging
Root cause analysis
Log investigation
Production troubleshooting

Example prompts:

"Investigate why API latency increased in the last hour"
"Find the root cause of the 500 errors in production"
"Analyze error patterns in the logs dataset"

The SRE skill learns from every debugging session. The memory system initializes automatically on first use. Memory persists at ~/.config/amp/memory/axiom-sre/ (global) or .agents/memory/axiom-sre/ (project/local).

Tell the agent to remember things

Remember this for next time.
Save this query, it worked.
Add to memory: the orders team uses #orders-oncall.

Learn automatically

The SRE skill automatically captures what works and what doesn't when:

A query or approach finds the root cause.
You correct it, and it records what didn't work and what did.
A debugging session completes successfully.

Customize with your own knowledge

Customize the SRE skill by adding your own knowledge to the knowledge base files:

kb/facts.md: Team contacts, Slack channels, conventions.
kb/integrations.md: Database connections, API endpoints.
kb/patterns.md: Failure patterns you've seen before.