Verification

A repository document, rendered for the site. View source markdown.

What this command does. cub installer is a released, open-source plugin for the cub CLI. cub installer setup pulls a catalog package and writes its Kubernetes files locally. It does not apply those files to a cluster; use kubectl, Argo CD, or Flux for delivery. The generated scripts stop before doing any work when the plugin or kustomize is missing.

New to cub? Install the cub CLI first. You can pull and render public catalog packages without an account. Commands that save or change ConfigHub data require you to sign in.

Generated at: 2026-07-30T12:38:02.000Z UTC · source: committed helm-expt evidence for this rendered repository document.

UNOFFICIAL/EXPERIMENTAL

Verification is how a reader checks a claim instead of trusting a screenshot. The npm commands in this repo are proof checks, not product install commands. Use the narrowest check that answers the question in front of you.

The Short Rule

Recipe, Render, Record, Route

The verification model has four steps:

StepMeaningWhat gets checked
RecipeLock and record the inputs as the recipe: chart source, version, values profile, release name, namespace, capability profile, and source lock.The inputs are pinned and recorded, so the render is reproducible.
RenderTurn that recipe into exact Kubernetes objects.The rendered object set matches the recorded package or Helm baseline.
RecordKeep the output evidence with the render: objects, diffs, receipts, scans, and observations.A later reviewer can see what changed and rerun the same proof boundary.
RouteName everything Helm leaves around the edges: hooks, CRDs, webhooks, generated facts, target prerequisites, Secrets, setup jobs, and GitOps handoff.Each extra is applied, observed, blocked, refused, or marked target-specific instead of being hidden inside prose.

This is why a flat YAML render is only the start. The value is that the render has recorded intent and named routes for the work that is not just a literal Kubernetes object.

Which Check Should I Run?

QuestionCommand or surfaceNeeds a cluster?What it proves
Are the generated website, docs, and data current?npm run site:verify, npm run docs:verify, npm run data:index:verifyNoGenerated surfaces match committed source and data.
Does a rendered Redis tutorial output match the catalog contract?npm run redis:verify-install:render -- ...NoA user's workdir render matches the expected chart/base/package contract.
Do I need the broad repo gate?npm run verifyNo cluster by defaultThe committed corpus, generated files, receipts, and docs are self-consistent.
Does regular Helm and cub reach the same live result?npm run kind-parity:run -- ...Yes, kindA fresh two-cluster Helm-vs-cub comparison for one chart/version/base.
Do committed two-cluster parity receipts still check out?npm run kind-parity:verifyNoExisting receipts and summaries are internally consistent.
Does the ConfigHub, OCI, and Argo path pass for a row?npm run live-parity:run -- ...Yes, kind plus live ConfigHub and OCI pathThe stricter live lane for a committed recipe/base.
What does a lane mean?npm run lane-tests:verify and Verification LanesNoThe lane matrix semantics and status vocabulary are still valid.
Can I validate a cub-scout receipt?cub-scout receipt validate <receipt.json>NoThe receipt fingerprint and structure validate locally.

User-Side And Maintainer Checks

User-side checks answer "can I inspect or compare this chart path myself?" Examples are cub installer setup, npm run redis:verify-install:render, and npm run kind-parity:run.

Maintainer checks answer "is the repo still internally consistent?" Examples are npm run site:verify, npm run docs:verify, npm run data:index:verify, and npm run verify.

Evidence-refresh checks answer "should we create new live receipts?" Examples are npm run live-parity:run, npm run outcomes:generate, and npm run status:dashboard. Run those deliberately because they can create or refresh evidence files.

Fresh Evidence And Committed Evidence

Committed evidence is already in the repo. It is good for review, publishing, and checking that generated pages still match the receipts and CSVs they cite.

Fresh evidence creates a new run. It may create kind clusters, use ConfigHub, publish OCI artifacts, wait for Argo or Flux, or write new receipts. Run fresh live lanes serially so clusters, namespaces, credentials, and receipts do not trample each other.

Subtopics

TopicUse it for
Verify It YourselfThe practical command list for offline checks, rendered installs, kind parity, ConfigHub/OCI live parity, and cub-scout receipts.
Verification LanesWhat each proof lane means, what it does not prove, and how to read pass, watch, blocked, missing, and fail.
Choosing CommandsWhen to use product commands versus repo verifiers.
Expected Results And ClustersWhat output should appear and which steps need a cluster.
Outcomes And TestsWhich repo promises map to which test commands and CSVs.
Live ParityHow to read live Helm-vs-ConfigHub parity status.
Chain Of ProofWhich boundary is proven by render receipts, ConfigHub receipts, delivery receipts, and live observations.
What We Refuse To ClaimThe refusal boundaries that keep proof language honest.
Two-Cluster Helm Parity HarnessThe stricter Helm-vs-cub kind harness.
NPM Test And Verification ScriptsThe full script catalog for maintainers.
Outcome CoverageGenerated lane coverage across the corpus.
Status DashboardCurrent generated proof counters and active queue.