Introducing kubevirt-ai-helpers — Claude Code plugins for KubeVirt development

[Lee Yarwood]

Hi all,

I'd like to share a project I've been working on: kubevirt-ai-helpers
[1], a collection of Claude Code plugins that automate common KubeVirt
development workflows.

This project is modeled after the openshift-eng/ai-helpers [2]
project, which provides similar plugins for OpenShift development
(Jira workflows, git operations, CI/CD, etc.). The two can be used
side by side — kubevirt-ai-helpers focuses on KubeVirt-specific
workflows while openshift-eng/ai-helpers covers the broader OpenShift
ecosystem.

If you use (or are curious about) Claude Code [3], these plugins
provide KubeVirt-specific commands for:

- Code review (/kubevirt:review) — review local branch changes against
KubeVirt best practices
- PR CI analysis (/kubevirt:review-ci) — analyze CI failures for a PR
with root cause identification and remediation suggestions
- CI health monitoring (/kubevirt:ci-health, /kubevirt:ci-lane,
/kubevirt:ci-search) — track CI stability, analyze job lanes, and
search for failure patterns
- CI triage and reporting (/kubevirt:ci-triage, /kubevirt:ci-report) —
AI-powered root cause analysis and stakeholder reports
- Linting (/kubevirt:lint) — run golangci-lint and generate a fix plan
with separate commits per linter
- VEP management (/kubevirt:vep-list, /kubevirt:vep-summary,
/kubevirt:vep-groom) — list, summarize, and review KubeVirt
Enhancement Proposals

The CI commands (ci-health, ci-lane, ci-search, ci-triage, ci-report)
are powered by my healthcheck [4] CLI tool, which provides live Prow
data crawling, aggregated ci-health data, and search capabilities via
search.ci.kubevirt.io. The AI plugins layer on top of healthcheck to
provide intelligent triage, root cause analysis, and reporting.

Installation is straightforward:

/plugin marketplace add lyarwood/kubevirt-ai-helpers
/plugin install kubevirt@kubevirt-ai-helpers

All commands are also available as Ambient Code Platform [5]
workflows, so you can use them without a local Claude Code
installation. When creating an Ambient session, attach a workflow from
this repository:

workflow_git_url: https://github.com/lyarwood/kubevirt-ai-helpers.git
workflow_path: workflows/kubevirt-review
workflow_branch: main

Available workflows: kubevirt-review, kubevirt-lint,
kubevirt-review-ci, kubevirt-vep-list, kubevirt-vep-summary,
kubevirt-vep-groom, kubevirt-ci-lane, kubevirt-ci-health,
kubevirt-ci-search, kubevirt-ci-triage, kubevirt-ci-report.

Contributions are very welcome, whether that's new commands,
improvements to existing ones, or bug reports. The repo has a
hello-world reference plugin and documentation on how to add commands.
If you have a repetitive workflow that could benefit from AI
assistance, consider turning it into a plugin.

There are no formal plans to move this under the kubevirt org at the
moment but I'm not against it if people think this project would be a
useful addition to the community.

Cheers,

Lee

[1] https://github.com/lyarwood/kubevirt-ai-helpers
[2] https://github.com/openshift-eng/ai-helpers
[3] https://docs.anthropic.com/en/docs/claude-code
[4] https://github.com/lyarwood/healthcheck
[5] https://github.com/ambient-code

[Felix Matouschek]

Lee, thanks for sharing this!

In the meanwhile I was able to add another plugin to kubevirt-ai-helpers: kubevirt:review-pr

It can be quite handy when reviewing PRs on GitHub.

Felix

[Lee Yarwood]

Hello all,

I wanted to follow up on the kubevirt-ai-helpers announcement and
highlight a few skills that are particularly useful in the lead-up to
VEP freeze. Whether you're authoring a new VEP, reviewing one, or
trying to get an overview of what's in flight, these commands can
help:

**`/kubevirt:vep-list`** -- List all open VEPs with filtering by SIG,
release, and tracking status. Useful for SIG leads getting a picture
of what's proposed for v1.9 and what still needs attention before
freeze.

**`/kubevirt:vep-summary <number>`** -- Get a quick TL;DR of any VEP:
motivation, design, current status, ownership, and progress across
Alpha/Beta/GA phases. Saves time when you need to catch up on a
proposal without reading the full document.

**`/kubevirt:vep-groom <pr-number>`** -- Automated review of a VEP
proposal PR against the template requirements and process guidelines.
Checks for missing sections, incomplete content, and gaps in project
tracking. Helpful for both authors (self-review before requesting
reviews) and reviewers (quick first pass).

**`/kubevirt:vep-find-reviewers <vep-number-or-sig>`** -- Identifies
potential reviewers by cross-referencing OWNERS_ALIASES with recent
git activity in the affected code areas. Surfaces who's actively
working in the relevant packages and flags OWNERS members who may be
less active. Useful when your VEP spans areas you're less familiar
with.

**`/kubevirt:vep-review-list`** -- Shows VEP proposal and
implementation PRs assigned to you for review, filtered to open PRs
only. If you're a VEP reviewer, this is a quick way to check what's
waiting on you as the freeze approaches.

These all work through Claude Code with the kubevirt plugin installed.
Setup instructions are in the [kubevirt-ai-helpers
README](https://github.com/lyarwood/kubevirt-ai-helpers).

If you're working on a VEP for v1.9, now is the time to get proposals
up and reviewers assigned ahead of freeze.

Cheers,

Lee