All products

Write an operations runbook with AI

Document a tested procedure in execution order. Give the operator prerequisites, risk labels, recovery steps and an observable check at the end.

Write an operations runbook with AI when the person covering the next shift needs to repeat a procedure reliably. The writing-runbooks skill takes the sequence your team knows and records the actor, required access and expected result for each step. A brief path at the top helps an experienced operator find their place.

Consider replacing a published file. Removing the old copy before the new one is available can leave readers stranded. Putting that consequence beside the step makes the order easier to follow under pressure. It also helps a reviewer spot where a recovery route has been assumed.

What it does

  • List the required versions and how to check them.
  • Mark actions that change production or need a manual step elsewhere.
  • Describe how to reverse a change, or record the consequences if it is irreversible.
  • Start troubleshooting entries with the symptom an operator can search for.
  • Map old settings to their replacements in a migration guide.

Install and use

  1. Install the maker’s technical-writer bundle, or this skill plus its required technical-writing companion. Provide an authorised procedure and the checks already performed.
  2. Ask for a runbook that labels each state-changing step and records what the operator should see afterwards.
  3. Have someone familiar with the process check the order and evidence. Label untested steps as draft.

FAQ

Source

Made by riekelt. Original skill: writing-runbooks, published on GitHub under the MIT licence.

View on GitHub