# Eks Operation Review

> >-

- **Type:** Skill
- **Install:** `agentstack add skill-aws-samples-sample-eks-operation-review-skill-devopsagent`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [aws-samples](https://agentstack.voostack.com/s/aws-samples)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT-0
- **Upstream author:** [aws-samples](https://github.com/aws-samples)
- **Source:** https://github.com/aws-samples/sample-eks-operation-review-skill/tree/main/DevOpsAgent
- **Website:** https://github.com/aws-samples/sample-eks-operation-review-skill

## Install

```sh
agentstack add skill-aws-samples-sample-eks-operation-review-skill-devopsagent
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# EKS Operation Review

This skill performs a structured 10-section operational assessment of a live EKS cluster, producing a rated report with prioritized recommendations.

## When to Activate

Activate for any request to:
- Audit, review, health-check, or score an EKS cluster's operational posture
- Assess a specific operational area (e.g., "check my EKS networking", "review RBAC on my cluster", "EKS observability assessment")
- Generate an EKS operational review report

**Do NOT activate for:** upgrade readiness assessments, cluster discovery, architectural design advice, general Kubernetes questions, AWS troubleshooting, cluster creation, or one-off kubectl commands.

## Execution Model

This skill runs **autonomously** — it does NOT pause for user input mid-execution. It performs a **HARD STOP** on ambiguity instead of guessing.

### HARD STOP Conditions

If any of these conditions are met, produce a HARD STOP output and no report:

```
## HARD STOP
- **Reason:** [why execution cannot proceed]
- **What was found:** [clusters/resources discovered]
- **What is needed:** [specific input required to proceed]
```

HARD STOP triggers:
1. Multiple EKS clusters found and none specified by the user — list them and stop
2. No EKS clusters found in the configured region
3. Cluster access denied — IAM permissions or Kubernetes RBAC (the agent's role cannot describe the cluster or reach the Kubernetes API)
4. EKS or Kubernetes API access is not available or does not respond

### Passing Inputs

For a clean single-pass run, specify: cluster name, region (if not default), and optionally which sections to assess. Example: "Run EKS operation review for cluster `prod-app` in `us-west-2`"

## Prerequisites

The DevOps Agent must have:
- **EKS cluster access** via its IAM role — see [EKS access setup](https://docs.aws.amazon.com/devopsagent/latest/userguide/configuring-integrations-and-knowledge-aws-eks-access-setup.html)
- **AWS API permissions** for: `eks:Describe*`, `eks:List*`, `ec2:DescribeSubnets`, `ec2:DescribeVpcs`, `iam:ListAttachedRolePolicies`, `iam:ListRolePolicies`, `logs:DescribeLogGroups`, `cloudwatch:DescribeAlarms`
- **Kubernetes RBAC access** to list/get Nodes, Pods, Deployments, Services, DaemonSets, Namespaces, and related resources

## Steering File Map

Before executing checks for any section, load the corresponding reference file from `references/`.

| User Request | Reference File(s) to Load |
|---|---|
| Full review / assess / audit / health check | ALL files in order: cluster-lifecycle → addon-management, then report-generation |
| Upgrade / version / deprecated API | `references/cluster-lifecycle.md` |
| IRSA / RBAC / access / pod identity / endpoint | `references/access-identity.md` |
| Logging / metrics / alerting / observability | `references/observability.md` |
| Resource requests / probes / PDB / image tags / storage | `references/workload-configuration.md` |
| IP / subnet / DNS / CoreDNS / network policy | `references/networking.md` |
| Autoscaling / Karpenter / HPA / topology spread | `references/autoscaling.md` |
| Deployment / rollout / CI/CD / graceful shutdown | `references/deployment-practices.md` |
| Runbook / on-call / backup / DR / Velero | `references/operational-processes.md` |
| Add-on / node monitoring / cluster insights | `references/addon-management.md` |
| Generate / write report | `references/report-generation.md` |
| IaC / GitOps / ArgoCD / Flux / drift | `references/infrastructure-as-code.md` |

## Assessment Overview

| # | Section | Key Checks |
|---|---------|------------|
| 01 | Cluster Lifecycle & Upgrades | Version currency, data plane alignment, deprecated APIs, add-on compatibility, upgrade process |
| 02 | Infrastructure as Code & GitOps | IaC provenance, GitOps tools, drift detection, RBAC in code |
| 03 | Access & Identity | IRSA/Pod Identity, least privilege RBAC, API server endpoint security, Pod Security Admission |
| 04 | Observability | Control plane logging, metrics stack, log aggregation, alerting |
| 05 | Workload Configuration | Resource requests/limits, health probes, PDBs, image tags, storage |
| 06 | Networking | IP capacity, CoreDNS health, network policies |
| 07 | Autoscaling | Cluster autoscaler/Karpenter, HPA, topology spread |
| 08 | Deployment Practices | Rollout strategy, CI/CD, graceful shutdown |
| 09 | Operational Processes | Backup/DR, tool presence (Velero, AWS Backup) |
| 10 | Add-on Management | Managed add-ons, node health monitoring, cluster insights |

~70-75% of items are fully automatable. Items requiring human knowledge (runbooks, on-call processes) are marked UNKNOWN with suggestions for investigation.

## Assessment Workflow

### Step 0: Pre-flight

Verify access before starting the assessment:

1. **List clusters** — use EKS ListClusters API. If multiple found and none specified → HARD STOP with list. If one found → proceed with that cluster.
2. **Describe cluster** — use EKS DescribeCluster API. Report: cluster name, Kubernetes version, platform version, region, status, authentication mode.
3. **Verify Kubernetes access** — list Nodes via Kubernetes API.
   - Success → proceed
   - Failure → HARD STOP with access error details

### Steps 1-10: Run Assessment

Load each reference file in section order. For each section:
1. Load the reference file from `references/`
2. Execute the checks described in it using available EKS and Kubernetes APIs
3. Rate each item using the rubric below

**Error recovery:** If a section fails entirely (API unreachable, permissions denied for all checks), mark all items as UNKNOWN with failure reason, then proceed to next section. Do not let one failed section block the entire assessment.

### Step 11: Generate Report

Load `references/report-generation.md` and produce the report following its structure and consistency rules.

## Rating Rubric

| Rating | Meaning |
|--------|---------|
| GREEN | Fully implemented — matches EKS best practices |
| AMBER | Partial or inconsistent — improvement opportunity |
| RED | Not implemented or significant gap — action needed |
| UNKNOWN | Cannot be determined from cluster data — investigate manually |

### Rules

- Only rate based on what was actually observed — never assume
- If a check fails or returns no data, mark UNKNOWN
- If a check fails with a permissions error, mark UNKNOWN and note the failure reason; in the report, direct the user to the "Resource Access (Prerequisites)" section of the skill README to grant the review read permissions
- Prioritize by blast radius: security > availability > cost
- Every RED finding must have a specific, actionable recommendation

## Report Output

### Filename

`EKS-Operation-Review---.md`

### Report Template

```markdown
# EKS Operation Review Report
Cluster: [name] | Region: [region] | Version: [version]
Date: [YYYY-MM-DD HH:MM]

## Executive Summary
[2-3 paragraphs. Strengths first, then gaps. Every rating mentioned must match findings.]

## Maturity Score
| Rating | Count | Percentage |
|--------|-------|------------|
| GREEN | X | X% |
| AMBER | X | X% |
| RED | X | X% |
| UNKNOWN | X | -- |

## Findings
[One table per section with columns: Item | Status | Current State | Recommendation | References]

## Prioritized Actions

### Critical (Address within 30 days)
[All RED items, ordered: security > availability > cost]

### Important (Address within 90 days)
[All AMBER items]

### Quick Wins
[Items fixable in < 1 hour]

## Items to Investigate Manually
[UNKNOWN items with specific questions]

## AWS Reference Links
[Grouped by topic — use pre-verified URLs from references/report-generation.md]
```

### Consistency Rules (MANDATORY)

1. Ratings must be consistent across the entire report
2. Prioritized Actions must reference the finding ID (e.g., "4.1 — Control Plane Logging RED")
3. Every RED must appear in Critical or Important; every AMBER in Important or Quick Wins
4. Executive Summary must match the findings — do not call something "critical" if it's AMBER
5. One row per finding in Prioritized Actions — never bundle multiple findings

### Report Footer

Append at the end:

```markdown
---
*This report was generated by an AWS DevOps Agent skill provided as sample code for
educational and demonstration purposes only. Findings should be reviewed and validated
before acting on them. See the project's README and LICENSE for full terms.*
```

## Live Data Caveat

This skill includes embedded reference tables (EKS version calendar, compatibility data, pre-verified AWS documentation URLs). These may become stale over time. Before rating version currency, cross-check against the official [EKS version calendar](https://docs.aws.amazon.com/eks/latest/userguide/kubernetes-versions.html) if documentation search is available in your Agent Space.

If live documentation lookup is not available, flag results as: "⚠️ Version data based on embedded reference table (last verified 2026-04-24). Verify against official docs."

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [aws-samples](https://github.com/aws-samples)
- **Source:** [aws-samples/sample-eks-operation-review-skill](https://github.com/aws-samples/sample-eks-operation-review-skill)
- **License:** MIT-0
- **Homepage:** https://github.com/aws-samples/sample-eks-operation-review-skill

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-aws-samples-sample-eks-operation-review-skill-devopsagent
- Seller: https://agentstack.voostack.com/s/aws-samples
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.