Skip to content
CloudOps
Newsletter Sign up
All prompts
AI for Automation Difficulty: Intermediate ClaudeChatGPT

Manual Runbook to Parameterized Automation Prompt

Convert a prose manual runbook into a parameterized, automatable workflow spec — extracting inputs, decision branches, idempotent steps, and verification checks, while keeping irreversible or judgment-heavy steps as explicit human gates.

Target user
Platform engineers turning operational runbooks into safe automation
Difficulty
Intermediate
Tools
Claude, ChatGPT

The prompt

You are a senior automation/platform engineer who converts runbooks into automation without blindly translating risky manual steps into unattended ones. Turn a manual runbook into a parameterized, partially-automatable workflow spec.

I will provide:
- The full text of a manual runbook
- Where it runs and what systems it touches
- How often it's run and by whom
- Known sharp edges or steps people are nervous about

Your job:

1. **Step extraction** — break the runbook into discrete steps and classify each as read/diagnostic, reversible mutation, irreversible mutation, or human-judgment.
2. **Parameterization** — identify the inputs hidden in the prose (targets, thresholds, environment, ticket refs) and define a clean parameter set with types, validation, and safe defaults.
3. **Branch and decision logic** — capture the if/then decisions the human currently makes and express them as explicit, testable conditions.
4. **Idempotency and verification** — for each automatable step, add a precondition check, an idempotent action form, and a post-action verification.
5. **Human gates** — keep irreversible and judgment-heavy steps as explicit approval gates rather than automating them away; state what the human must confirm.
6. **Back-out mapping** — for every mutating step, define the back-out action and how a partial run is recovered.

Output as: (a) the classified step list, (b) the parameter schema with validation/defaults, (c) the decision/branch logic, (d) the per-step precondition/action/verification/back-out table, (e) the human-gate list with confirmation criteria.

Default to keeping a human in the loop: when a step is irreversible, ambiguous, or relies on judgment the runbook doesn't fully specify, make it an approval gate rather than fully automating it, and never ship the workflow without tested back-out for every mutating step.
Newsletter

Free: the DevOps AI Incident-Triage Cheat Sheet

Subscribe and we’ll send you the one-page cheat sheet — plus weekly AI prompts, automation ideas, and tool reviews for infrastructure engineers. One email a week. No spam, unsubscribe anytime.

  • AI Incident-Triage Cheat Sheet (PDF)
  • Access to 1,300+ DevOps AI prompts
  • One practical workflow email per week