SRE Practice Skills
Eight Agent Skills that teach Claude Code and compatible agents SRE practices such as SLOs, burn-rate alerts, incident response and postmortems.
At a Glance
Apache-2.0 licensed set of eight agent skills for SRE practices, free to use, modify and distribute. The license applies only to the original repository content, not Google's books.
Engagement
Available On
Alternatives
Listed Oct 2026
About SRE Practice Skills
SRE Practice Skills is an open-source collection of eight Agent Skills published by paulip114 on GitHub. It turns practices from Google's two free SRE books into step-by-step procedures, checklists and templates that Claude Code and other Agent Skills-compatible agents load when a request matches. The project is unofficial and not affiliated with or endorsed by Google.
What It Is
The repository packages SRE guidance as standard SKILL.md skills rather than as an application. Each skill links to the chapters of Site Reliability Engineering and The Site Reliability Workbook it draws on, but the procedures are written in the author's own words and do not reproduce the books' text. The skills cover SLOs and error budgets, monitoring and alerting, incident response, postmortems, on-call and toil, release and launch, capacity and system design, and data reliability.
How It Works in Practice
An agent loads a skill automatically when the request matches its description, or you can name one directly, for example asking it to use sre-slos-and-error-budgets to propose SLOs for a checkout API. In the README's illustrative example, the skill guides the agent to produce starter SLO targets derived from observed performance, allowed bad events per window, known measurement blind spots, a filled-in SLO document and sign-off steps.
Templates and Tools
The skills ship with reference material: an SLI menu by component type, SLO document and error budget policy templates, Prometheus recording and alerting rules, a live incident document template, a postmortem template, a launch checklist, a production best-practices audit list and a back-of-envelope estimation cheatsheet. A Python script, burn_rate.py, uses only the standard library to calculate multiwindow burn-rate alert thresholds for a given SLO and period.
Setup Path
Install all skills with the skills CLI via npx skills add, install a single skill or target Claude Code only, or install globally. Alternatively, clone the repository and copy the skill folders into ~/.claude/skills. The README is provided in English and Traditional Chinese.
Community Discussions
Be the first to start a conversation about SRE Practice Skills
Share your experience with SRE Practice Skills, ask questions, or help others learn from your insights.
Pricing
Open Source
Apache-2.0 licensed set of eight agent skills for SRE practices, free to use, modify and distribute. The license applies only to the original repository content, not Google's books.
- Eight SRE agent skills
- Install via npx skills add paulip114/sre-practice-skills
- Templates and burn-rate calculator script
- Apache License 2.0 applies only to original content in this repository, not Google's books (CC BY-NC-ND 4.0)
Capabilities
Key Features
- Eight Agent Skills covering SRE practices
- SLO and error budget guidance with SLO document and policy templates
- Monitoring and alerting guidance including multiwindow multi-burn-rate alerts
- Prometheus recording and alerting rule references
- Burn-rate calculator script using only the Python standard library
- Incident response procedures and live incident document template
- Blameless postmortem template and review checklist
- On-call, toil, release, launch and production readiness guidance
- Capacity planning and system design cheatsheet
- Data reliability guidance for pipelines, backups and integrity
- Automatic skill loading when a request matches, or explicit invocation by name
- Install via skills CLI or manual copy
