Guide · Agent skill
How to use the Engineering Runbook Writer agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The eng-runbook skill is a set of instructions that teaches an AI coding agent how to write an engineering runbook for a service: an ops document with alerts, dashboard links, copy-paste procedures, an on-call rotation and an incident response checklist. This guide shows you how to install it, how to phrase a request so the agent triggers it, and what to check before you rely on the output. By the end you will be able to produce a first-draft runbook as a single HTML page.
Engineering Runbook Writernexu-io/open-design · Apache-2.0 Download the skill Quick start
- 1Download the ZIP. Get the eng-runbook skill ZIP from the design-templates/eng-runbook page. It contains SKILL.md and example.html.
- 2Install the skill. Unzip it into ~/.claude/skills/ for personal use, into .claude/skills/ for one project, or upload it under Settings → Capabilities → Skills in the Claude apps.
- 3Add a DESIGN.md file. Place a DESIGN.md design system file in your project so the skill can read it first.
- 4Write the service brief. Name the service and supply its owner, severity, alerts, dashboards and procedures.
- 5Ask for a runbook. Use a request containing runbook, ops doc, on-call or SRE doc, such as writing a runbook for your auth service.
- 6Review the HTML page. Open the generated single-page HTML and check every command against your real environment before use.
What the skill makes the agent do
Once the skill is installed, asking the agent for an ops doc, on-call guide, SRE doc or runbook for a service makes it follow a fixed layout. The layout has a header with service name, owner team, severity tier and version; a service summary with a dependency list; an alerts table; a dashboards and links list; three to four common procedures with code blocks such as deploy, rollback and key rotation; an on-call rotation table; and a five-step incident response checklist.
The output is one self-contained HTML page with an inline style block. The skill first reads a DESIGN.md design system file in your project and styles the page to match it. The skill writes documentation only. It is not a monitoring tool and cannot check real alerts, dashboards or systems.
Install the eng-runbook skill
Download the ZIP for the skill from the design-templates/eng-runbook page on our site. The skill folder contains two files: SKILL.md and example.html.
To install for yourself across projects, unzip it into ~/.claude/skills/. To install for one project only, unzip it into that project's .claude/skills/ folder. Alternatively, in the Claude apps, open Settings → Capabilities → Skills and upload the skill there.
- Personal install: ~/.claude/skills/
- Project install: .claude/skills/ inside your project
- Claude apps: Settings → Capabilities → Skills
Prepare your project before you ask
The skill requires a DESIGN.md design system file in the project. The workflow reads it first, so add one before you make the request. If the file is missing, the output will not be styled to your design system.
Have the service details ready in the brief. The skill identifies the service from the brief, so name it and include the facts you know: owner, severity, alerts, dashboards and the procedures you actually use. The skill does not discover these for you.
Trigger the skill with natural requests
The skill activates when your request mentions terms such as runbook, ops doc, on-call, SRE doc, service runbook or 运维手册. You do not need to type the skill name. Phrase the request around the service you want documented.
The skill's own example prompt is a good model for the shape of a request.
- Write a runbook for our auth service — alerts, dashboards, common procedures, on-call rotation.
- Create an on-call guide for the billing API with its alerts and an incident checklist.
- Draft an SRE doc for our search service, styled to DESIGN.md.
Review the output before anyone uses it
Treat the result as a first draft. The common procedures such as deploy, rollback and key rotation are written by the model and may be wrong for your environment. Check every command against your real tooling before anyone runs it, especially in production.
Do not use the skill for live incident response. It cannot read your real alerts or systems, so the content reflects only what you supplied in the brief.
- Verify each command, path and hostname against your environment.
- Replace placeholder owners, dashboards and rotation names with real ones.
- Confirm the escalation and incident steps match your team's process.
Safety notes: scripts and network
The eng-runbook skill does not run scripts and does not use the network, according to its listing. It produces a document only. Because it writes commands for you to copy, the main risk is that a command looks plausible but is wrong for your system. Review it as you would any command from a colleague who has not seen your infrastructure.
What people use it for
- A first-draft runbook for a new service that the team has not documented yet.
- Standardising the structure of on-call documentation across several services.
- Laying out existing service details consistently in one styled HTML page.
- Producing a rollback and key-rotation procedure section to be checked by an SRE before publication.
Frequently asked questions
How do I install the eng-runbook skill?
Download the ZIP from the design-templates/eng-runbook page on our site. Unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder, or upload it in the Claude apps under Settings → Capabilities → Skills.
What does the eng-runbook skill require?
It requires a DESIGN.md design system file in the project, because the workflow reads it first. It does not run scripts and does not use the network.
Is the eng-runbook skill free?
The skill is listed under the Apache-2.0 licence. Check the licence file in the source repository for the terms that apply to your use.
Can the skill monitor my live alerts or dashboards?
No. It is a template for writing documentation, not a monitoring tool. It writes a document and cannot check real alerts, dashboards or systems.
Are the generated deploy and rollback commands safe to run?
Not without review. The procedures are written by the model and may be wrong for your environment, so verify each command against your own tooling before running it, particularly in production.
What is the difference between a runbook and an on-call guide here?
The skill treats them as the same kind of document. Requests mentioning runbook, ops doc, on-call guide or SRE doc all produce the same fixed layout with alerts, procedures, rotation and incident checklist.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Choose eng-runbook if you need a consistently structured first draft of a service runbook quickly and already know the service details. Pick something else if you need live monitoring or incident response, since this skill only writes documentation and cannot check real systems.
Download the skill