---
name: kubernetes-deployment-documenter
description: "Turns a Deployment manifest into a one-page reference an SRE or developer can read in two minutes: what runs, how it scales, what it needs, and what to watch during a rollout. Use when a Kubernetes Deployment manifest needs a readable reference page for SREs or developers who should not have to parse the YAML. Do not use for validating a manifest for errors or rewriting it to best practices. Related topics: coding, kubernetes."
---

# Kubernetes Deployment Documenter

Turns a Deployment manifest into a one-page reference an SRE or developer can read in two minutes: what runs, how it scales, what it needs, and what to watch during a rollout.

## When to use

Use this skill when a Kubernetes Deployment manifest needs a readable reference page for SREs or developers who should not have to parse the YAML. Do not use it for validating a manifest for errors or rewriting it to best practices.

## Workflow

1. Ask the user for each input listed below in a single message, skipping any they have already given. Where options are listed, offer them, but accept other answers too.
2. Extract every fact from the manifest a reader needs: identity, selector, replicas, update strategy, containers, ports, probes, volumes, and service account, with nothing invented.
3. Write the Overview in plain language pitched at the chosen audience, so it answers what the workload does and how it scales in under five sentences.
4. Fill the Spec Summary and Containers tables from the extracted facts, one row per container with image, tag, ports, and probes.
5. Add the Resources and Environment Variables sections only when toggled on, masking secret values and naming their source instead.
6. List surprising settings in Notes, such as a latest tag or a missing probe, and deliver the document in the requested format with the sections in order.

## Inputs

- **Deployment manifest** (longer free text)
- **Intended audience** (choose one: SRE, Developer, Architect)
- **Detail level** (choose one: High, Medium, Low)
- **Include resource limits** (choose one: Yes, No)
- **Include environment variables** (choose one: Yes, No)
- **Output format** (choose one: Markdown, HTML, AsciiDoc)

## Output

A single document in the selected format with sections in this order: Overview, Spec Summary, Containers table, Resources (optional), Environment Variables (optional), Notes.

## Prompt template

Fill the inputs into this template and follow it as the full instruction.

### Kubernetes Deployment Documenter

Document my Kubernetes Deployment so the team can understand it without reading the YAML.

- Deployment manifest: {{Deployment manifest}}
- Intended audience: {{Intended audience}}
- Detail level: {{Detail level}}
- Include resource limits: {{Include resource limits}}
- Include environment variables: {{Include environment variables}}
- Output format: {{Output format}}

Read the manifest before writing a word and pull out the facts a reader needs: name, namespace, labels and selector, replica count and update strategy, each container with its image and tag, ports, probes, volumes, and the service account. Write the Overview in plain language for the chosen audience: an SRE wants rollout behavior and failure modes, a developer wants ports and configuration, an architect wants how it fits with the rest of the system. Put the facts in tables rather than prose, and scale the depth to the detail level. Mask secret values and note where they come from (Secret or ConfigMap reference) rather than echoing them. Add a short Notes section for anything surprising: a latest tag, a missing probe, no resource limits, a single replica with RollingUpdate. Return one document in the selected format with these sections in order: Overview, Spec Summary, Containers, Resources (when included), Environment Variables (when included), Notes.

## Pitfalls

- Copying secret values or base64 blobs from env entries into the document, which turns a reference page into a leak.
- Describing what the Deployment should do instead of what the manifest actually says, so the doc and the cluster drift apart.
- Burying the replica count and update strategy in prose where an SRE on call cannot find them.
