awesome-sre-skills

A curated list of AI-SRE skills, MCP servers, and resources for building autonomous reliability into engineering organizations.

Maintained by Anyshift. We're building the AI-SRE skills category and tracking what good looks like here.

Contents

• Skill libraries
• MCP servers for SRE
• Reading
• Related awesome lists
• Contributing

Skill libraries

Methodology-shaped skills for AI agents working on reliability: incident investigation, change-impact analysis, oncall handover, postmortem authoring, reliability audits.

• anthropics/skills - Anthropic's reference repository of Agent Skills using the SKILL.md folder format that agents discover and load on demand.
• anthropics/claude-cookbooks - Anthropic's collection of notebooks and recipes for agentic workflows and tool use that reliability agents can build on.
• anyshift-io/sre-skills - Apache 2.0 library of methodology-shaped SRE skills published by Anyshift.
• openai/skills - OpenAI's reference collection of agent skills with operational examples.

MCP servers for SRE

Model Context Protocol servers that give AI agents access to the systems oncall engineers actually use.

• Datadog MCP Server - Datadog's official remote MCP server exposing logs, metrics, traces, monitors, and incidents to AI clients.
• GitHub MCP Server - GitHub's official MCP server giving agents access to repositories, issues, pull requests, and Actions for change context.
• Grafana MCP Server - Grafana's official MCP server exposing dashboards, Prometheus and Loki queries, and alerting to AI agents.
• incident.io Remote MCP Server - incident.io's official remote MCP server for querying incidents, alerts, on-call schedules, and post-mortems.
• Kubernetes MCP Server - An MCP server for Kubernetes and OpenShift that lets agents inspect and manage cluster resources.
• modelcontextprotocol/servers - The official reference and community index of MCP servers for connecting agents to operational systems.
• PagerDuty MCP Server - PagerDuty's official MCP server exposing incidents, services, and on-call schedules to MCP clients.
• Rootly MCP Server - Rootly AI Labs MCP server exposing incident-management tools and example incident-responder skills.
• Sentry MCP Server - Sentry's official MCP server for querying errors, issues, and performance data during investigations.
• tessl.io/registry - Searchable registry of MCP servers covering observability, ticketing, and adjacent SRE tooling.

Reading

Curated reading that shaped how the AI-SRE category thinks.

• Equipping agents for the real world with Agent Skills - Anthropic's engineering write-up on the Agent Skills format and the progressive-disclosure design behind skill libraries.
• How Complex Systems Fail - Richard Cook's 18-point treatise on failure in complex systems, a core reference for incident analysis.
• Site Reliability Engineering (Google SRE Book) - Google's foundational SRE book, free online, defining error budgets, SLOs, and on-call practice.
• The Site Reliability Workbook - Google's companion volume with practical SRE implementation guidance, free online.

Related awesome lists

• awesome-ai-sre - A curated list of AI-powered tools, agents, and resources for Site Reliability Engineering.
• awesome-AIOps - A curated list of academic research and industrial materials on AI for IT operations.
• awesome-incident-response - A curated list of tools and resources for incident response.
• awesome-observability - A curated list of observability tools, platforms, and learning resources.
• awesome-sre - Dastergon's curated list of Site Reliability and Production Engineering resources.
• awesome-sre-agents - Last9's curated list of AI-powered DevOps and SRE agents and tooling.
• awesome-sre-tools - Squadcast's curated list of Site Reliability and Production Engineering tools.
• howtheysre - A curated collection of publicly available resources on how organizations practice SRE.

Contributing

See CONTRIBUTING.md. Open a PR adding your entry to the right section with a one-sentence description.

License

Apache 2.0.