Changelog

All notable changes to this project will be documented in this file.

[0.7.7] - 2026-02-24

Bug Fixes

• Resolve gosec lint issues and bump golangci-lint to v2.10.1 by @mchmarny
• Guard against empty path in NewFileReader after filepath.Clean by @mchmarny
• Pass cluster K8s version to Helm SDK chart rendering by @mchmarny
• (e2e) Update deploy-agent test for current snapshot CLI by @mchmarny
• Prevent snapshot agent Job from nesting agent deployment by @mchmarny

Build

• Release v0.7.7 by @mchmarny

CI/CD

• Harden workflows and reduce duplication by @mchmarny

Features

• (ci) Add metrics-driven cluster autoscaling validation with Karpenter + KWOK by @dims
• (validator) Add Go-based CNCF AI conformance checks by @dims
• (validator) Self-contained DRA conformance check with EKS overlays by @dims
• (validator) Self-contained gang scheduling conformance check by @dims
• (validator) Upgrade conformance checks from static to behavioral validation by @dims
• Add conformance evidence renderer and fix check false-positives by @dims
• (validator) Replace helm CLI subprocess with Helm Go SDK for chart rendering by @xdu31
• Add HPA pod autoscaling evidence for CNCF AI Conformance by @yuanchen8911
• (collector) Add Helm release and ArgoCD Application collectors by @mchmarny
• Add cluster autoscaling evidence for CNCF AI Conformance by @yuanchen8911
• (ci) Binary attestation with SLSA Build Provenance v1 by @lockwobr

Tasks

• (ci) Remove redundant DRA test steps from inference workflow by @dims
• Upgrade Go to 1.26.0 by @mchmarny
• (validator) Remove Job-based checks from readiness phase, keep constraint-only gate by @xdu31
• (recipe) Add conformance recipe invariant tests by @dims

[0.7.7] - 2026-02-24

Bug Fixes

• Resolve gosec lint issues and bump golangci-lint to v2.10.1 by @mchmarny
• Guard against empty path in NewFileReader after filepath.Clean by @mchmarny
• Pass cluster K8s version to Helm SDK chart rendering by @mchmarny
• (e2e) Update deploy-agent test for current snapshot CLI by @mchmarny
• Prevent snapshot agent Job from nesting agent deployment by @mchmarny

CI/CD

• Harden workflows and reduce duplication by @mchmarny

Features

• (ci) Add metrics-driven cluster autoscaling validation with Karpenter + KWOK by @dims
• (validator) Add Go-based CNCF AI conformance checks by @dims
• (validator) Self-contained DRA conformance check with EKS overlays by @dims
• (validator) Self-contained gang scheduling conformance check by @dims
• (validator) Upgrade conformance checks from static to behavioral validation by @dims
• Add conformance evidence renderer and fix check false-positives by @dims
• (validator) Replace helm CLI subprocess with Helm Go SDK for chart rendering by @xdu31
• Add HPA pod autoscaling evidence for CNCF AI Conformance by @yuanchen8911
• (collector) Add Helm release and ArgoCD Application collectors by @mchmarny
• Add cluster autoscaling evidence for CNCF AI Conformance by @yuanchen8911

Tasks

• (recipe) Add conformance recipe invariant tests by @dims
• (validator) Remove Job-based checks from readiness phase, keep constraint-only gate by @xdu31
• (ci) Remove redundant DRA test steps from inference workflow by @dims
• Upgrade Go to 1.26.0 by @mchmarny

[0.7.6] - 2026-02-21

Tasks

• Codebase consistency fixes and test coverage by @mchmarny
• Rename cleanup by @mchmarny
• Remove redundant local e2e script by @mchmarny
• Remove flox environment support by @mchmarny
• Remove empty .envrc stub by @mchmarny

[0.7.5] - 2026-02-21

Bug Fixes

• (ci) Add packages:read permission to deploy job by @mchmarny

[0.7.4] - 2026-02-21

Bug Fixes

• (ci) Re-enable CDI for H100 kind smoke test by @dims
• Update inference stack versions and enable Grove for dynamo workloads by @yuanchen8911
• (ci) Harden workflows and improve CI/CD hygiene by @mchmarny
• (ci) Use pull_request_target for write-permission workflows by @mchmarny
• (ci) Break long lines in welcome workflow to pass yamllint by @dims
• Remove admission.cdi from kai-scheduler values by @yuanchen8911
• (ci) Add pull_request trigger to vuln-scan workflow by @mchmarny
• Enable DCGM exporter ServiceMonitor for Prometheus scraping by @yuanchen8911
• (ci) Combine path and size label workflows to prevent race condition by @yuanchen8911
• Add markdown rendering to chat UI and update CUJ2 documentation by @yuanchen8911
• Add kube-prometheus-stack as gpu-operator dependency by @yuanchen8911
• Skip --wait for KAI scheduler in deploy script by @yuanchen8911
• (ci) Lower vuln scan threshold to MEDIUM and add container image scanning by @dims
• (docs) Update bundle commands with correct tolerations in CUJ demos by @yuanchen8911
• (ci) Run attestation and vuln scan concurrently in release workflow by @dims
• Remove trailing quote from skyhook no-op package version by @yuanchen8911
• Remove nodeSelector from EBS CSI node DaemonSet scheduling by @yuanchen8911
• Move DRA controller nodeAffinity override to EKS overlay by @yuanchen8911
• (ci) Use PR number in KWOK concurrency group by @mchmarny

Features

• (ci) Add OSS community automation workflows by @mchmarny
• Add CUJ2 inference demo chat UI and update CUJ2 instructions by @yuanchen8911
• Add DRA and gang scheduling test manifests for CNCF AI conformance by @yuanchen8911
• (ci) Collect AI conformance evidence in H100 smoke test by @dims
• (ci) Add DRA GPU allocation test to H100 smoke test by @dims
• Add expected-resources deployment check for validating Kubernetes resources exist by @xdu31
• Add CNCF AI Conformance evidence collection by @yuanchen8911
• (skyhook) Temporarily remove skyhook tuning due to bugs by @ayuskauskas
• Add GPU training CI workflow with gang scheduling test by @dims
• (ci) Add CNCF AI conformance validations to inference workflow by @dims
• (ci) Add HPA pod autoscaling validation to inference workflow by @dims
• (ci) Add ClamAV malware scanning GitHub Action by @dims
• Add two-phase expected resource auto-discovery to validator by @xdu31
• Add support for workload-gate and workload-selector by @ayuskauskas

Refactor

• Move examples/demos to project root demos directory by @mchmarny
• Move kai-scheduler and DRA driver to base overlay for CNCF AI conformance by @yuanchen8911
• Rename PreDeployment to Readiness across codebase and docs by @xdu31

Tasks

• Update demos by @mchmarny
• Update s3c demo by @mchmarny
• Update demos by @mchmarny
• Update e2e demo by @mchmarny
• Update e2e demo by @mchmarny
• Update e2e demo by @mchmarny
• Update e2e demo by @mchmarny
• Improve consistency across GPU CI workflows by @dims
• Update cuj1 by @mchmarny

[0.7.3] - 2026-02-18

Bug Fixes

• Add merge logic for ExpectedResources, Cleanup, and ValidationConfig in recipe overlays by @xdu31

[0.7.2] - 2026-02-18

Bug Fixes

• Pipe test binary output through test2json for JSON events by @mchmarny

[0.7.1] - 2026-02-18

Bug Fixes

• Enable GPU resources and upgrade DRA driver to 25.12.0 by @yuanchen8911

Features

• Add test isolation to prevent production cluster access by @mchmarny
• Multi-stage Dockerfile.validator with CUDA runtime base by @mchmarny

Refactor

• (phase1) Fix best practice violations by @mchmarny
• (phase2) Extract duplicated code to pkg/k8s/pod by @mchmarny
• (phase3) Optimize Kubernetes API access and simplify HTTPReader by @mchmarny
• (phase4) Polish codebase with cleanup and TODO resolution by @mchmarny

Tasks

• Clean up change log by @mchmarny
• Cleanup docker file by @mchmarny

[0.7.0] - 2026-02-18

Bug Fixes

• Remove fullnameOverride from dynamo-platform values by @yuanchen8911
• Disable CDI in GPU Operator for dynamo inference recipes by @yuanchen8911

Features

• (ci) Add Dynamo vLLM smoke test and fix etcd/NATS naming by @dims
• Feat/adding smi test by @iamkhaledh, @jaydu

[0.6.4] - 2026-02-17

Bug Fixes

• Default validation-namespace to namespace when not explicitly set by @mchmarny
• Build aicr CLI in validator image and update binary path by @mchmarny

Refactor

• (ci) Decompose gpu-smoke-test into composable actions by @dims

Tasks

• Correct test command prior to PR by @mchmarny
• Clean changelog by @mchmarny

[0.6.3] - 2026-02-17

Bug Fixes

• Wrap bare errors, add context timeouts, use structured logging by @mchmarny
• (ci) Deduplicate tools, add robustness and consistency improvements by @mchmarny
• (ci) Increase GPU Operator ClusterPolicy timeout to 10 minutes by @mchmarny
• (ci) Harden H100 smoke test workflow by @dims

Features

• (ci) Add CUJ2 inference workflow to H100 smoke test by @dims
• Add kind-inference overlays and chainsaw health checks by @dims
• Skyhook gb200 by @ayuskauskas
• Validator generator, add test coverage, wire image-pull-secret by @mchmarny

Refactor

• Remove dead code, fix perf hotspots, add test coverage by @mchmarny
• (ci) Extract gpu-cluster-setup action, let H100 deploy GPU operator via bundle by @dims
• Standardize kind values to PascalCase by @mchmarny

[0.6.2] - 2026-02-13

CI/CD

• Add actions:read permission to security-scan job by @mchmarny
• Eliminate hardcoded versions and consolidate CI workflows by @mchmarny
• Harden checkout credentials, add checksum verification, fail-fast off by @mchmarny
• Skip SBOM generation in packaging dry run by @mchmarny

Tasks

• Clean up changelog by @mchmarny

[0.6.1] - 2026-02-13

Features

• (skyhook-customizations) Use overrides and switch to nvidia_tuned by @ayuskauskas
• Vendor Gateway API Inference Extension CRDs (v1.3.0) by @yuanchen8911
• (test) Add standalone resource existence checker for ai-conformance by @dims

Bug Fixes

• Protect system namespaces from deletion in undeploy.sh by @yuanchen8911
• Rename skyhook CR to remove training suffix by @yuanchen8911
• Add nats storageClass for EKS dynamo deployment by @yuanchen8911
• Mount host /etc/os-release in privileged snapshot agent by @yuanchen8911

CI/CD

• Add GPU smoke test workflow using nvkind by @dims
• Enable copy-pr-bot by @dims
• Setup vendoring for golang by @lockwobr
• Deduplicate test jobs into reusable qualification workflow by @mchmarny

Tasks

• Exclude git from sandbox for GPG commit signing by @mchmarny
• Code quality cleanup across codebase by @mchmarny
• Rename skyhook customization manifest to remove training suffix by @yuanchen8911
• (recipe) Move embedded data to recipes/ at repo root by @lockwobr

[0.5.16] - 2026-02-12

Bug Fixes

• Use POSIX-compatible redirects in KWOK parallel test script by @yuanchen8911
• KubeFlow patches by @coffeepac

Features

• Add tools/describe for overlay composition visualization by @mchmarny
• Restructure inference overlay hierarchy by @yuanchen8911

[0.5.15] - 2026-02-11

Bug Fixes

• Use universal binary name for macOS in install script by @mchmarny
• Use per-arch darwin binaries instead of universal binary by @mchmarny

[0.5.14] - 2026-02-11

Bug Fixes

• Resolve EKS deployment issues for multiple components by @yuanchen8911
• Preserve version prefix in deploy.sh for helm install by @yuanchen8911

[0.5.13] - 2026-02-11

Features

• Implement Job-based validation framework with test wrapper infrastructure by @xdu31
• Add kai-scheduler component for gang scheduling by @yuanchen8911
• Add dynamo-platform and dynamo-crds for AI inference serving by @yuanchen8911
• Add kgateway for CNCF AI Conformance inference gateway by @yuanchen8911
• Add basic spec parsing by @cullenmcdermott
• Add undeploy.sh script to Helm bundle deployer by @mchmarny

Bug Fixes

• Helm-compatible manifest rendering and KWOK CI unification by @mchmarny
• Resolve staticcheck SA5011 and prealloc lint errors by @yuanchen8911
• Fix deploy.sh failing when run from within the bundle directory. by @yuanchen8911
• Use upstream default namespaces for components by @yuanchen8911
• Update kubeflow paths by @coffeepac

Tasks

• Split validator docker build into per-arch images with manifest list by @mchmarny

[0.4.1] - 2026-02-08

Bug Fixes

• Remove redundant driver resource limits by @yuanchen8911
• Make configmap for kernel module config a template; clean up unu… by @valcharry
• Re-enable cert-manager startupapicheck by @yuanchen8911
• Disable skyhook LimitRange by bumping to v0.12.0 by @yuanchen8911
• Set fullnameOverride to remove aicr-stack- prefix by @yuanchen8911
• Open webhook container ports in NetworkPolicy workaround by @yuanchen8911

Tasks

• Clean up changelog by @mchmarny
• Update installation instructions by @mchmarny
• Add validation to e2d demo by @mchmarny
• Add b200 snapshot and report by @mchmarny
• Update b200 snapshot by @mchmarny
• Disable scans until GHAS is enabled again by @mchmarny
• Disable upload until ghas is enabled by @mchmarny
• Remove duplicate code scan by @mchmarny
• Add license to b200 example by @mchmarny

[0.4.0] - 2026-02-06

Features

• Add aws-efa component by @Kevin-Hawkins
• Fix and improve ConfigMap and CR deployment by @yuanchen8911
• Skyhook, split customizations to their own component and add training by @ayuskauskas
• Add skeleton multi-phase validation framework by @xdu31
• Custom resources must explicitly set their helm hooks OR opt out by @ayuskauskas
• Enhance validate command with multi-phase and agent support by @mchmarny

Bug Fixes

• (e2e-test) Create snapshot namespace before RBAC resources by @yuanchen8911
• (tools) Make check-tools compatible with bash 3.x by @yuanchen8911
• Correct manifest path in external overlay example by @mchmarny
• Add NetworkPolicy workaround for nvsentinel metrics-access restriction by @yuanchen8911
• Disable aws-ebs-csi-driver by default on EKS by @yuanchen8911
• Prevent driver OOMKill during kernel module compilation by @yuanchen8911
• Update CDI configuration and DEVICE_LIST_STRATEGY for gpu-operator by @yuanchen8911

Tasks

• Rename platform pytorch to kubeflow and add kubeflow-trainer component by @mchmarny
• Reduce e2e test duplication and add CUJ1 coverage by @mchmarny
• Remove daily scan from blocking prs by @mchmarny
• Add cuj1 demo by @mchmarny

[0.3.3] - 2026-02-04

Tasks

• Adjust release commit message order by @mchmarny

[0.3.2] - 2026-02-04

Tasks

• Include non-conventional commits in changelog by @mchmarny
• Update release commit message format by @mchmarny

[0.3.1] - 2026-02-04

Features

• Add aws-efa component by @Kevin-Hawkins

Refactor

• Use structured errors and improve test coverage by @mchmarny

Tasks

• Remove daily scan from blocking prs by @mchmarny
• Add Claude instructions to not co-authored commits by @mchmarny
• Allow attribution but not co-authoring by @mchmarny
• Moved coauthoring into main claude doc by @mchmarny

[0.3.0] - 2026-02-04

Bug Fixes

• Add contents:read permission for coverage comment workflow by @dims
• Use /tmp paths for coverage artifacts by @dims
• Rename prometheus component to kube-prometheus-stack by @yuanchen8911
• Remove namespaceOverride from nvidia-dra-driver-gpu values by @yuanchen8911

CI/CD

• Add license verification workflow by @dims
• Add license verification workflow by @dims
• Add CodeQL security analysis workflow by @dims
• Use copy-pr-bot branch pattern for PR workflows by @dims
• Trigger workflows on branch create for copy-pr-bot by @dims
• Skip workflows on forks to prevent duplicate check runs by @dims
• Match nvsentinel workflow pattern for copy-pr-bot by @dims

Features

• Add coverage delta reporting for PRs by @dims
• Link GitHub usernames in changelog by @dims
• Add structured CLI exit codes for predictable scripting by @dims
• Add fullnameOverride to remove release prefix from deployment names by @yuanchen8911

Tasks

• Rename default claude file to follow convention by @mchmarny
• Add .claude/settings.local.json to ignore by @mchmarny
• Add copy-pr-bot configuration by @dims
• Refactor tools-check into standalone script by @mchmarny

[0.2.2] - 2026-02-01

Bug Fixes

• Preserve manual changelog edits during version bump by @mchmarny

[0.2.1] - 2026-02-01

Bug Fixes

• Use workflow_run for PR coverage comments on fork PRs by @dims
• Add actions:read permission for artifact download by @dims

Features

• Add contextcheck and depguard linters by @dims
• Add stale issue and PR automation by @dims
• Add Dependabot grouping for Kubernetes dependencies by @dims
• Add automatic changelog generation with git-cliff by @mchmarny

Tasks

• Add dims in maintainers by @mchmarny
• Add owners file by @mchmarny
• Fix code owners by @mchmarny
• Replace explicit list with a link to the maintainer team by @mchmarny
• Update code owners by @mchmarny

[0.2.0] - 2026-01-31

Bug Fixes

• Support private repo downloads in install script by @mchmarny
• Skip sudo when install directory is writable by @mchmarny

[0.1.5] - 2026-01-31

Bug Fixes

• Add GHCR authentication for image copy by @mchmarny

[0.1.4] - 2026-01-31

Features

• Add Artifact Registry for demo API server deployment by @mchmarny

[0.1.3] - 2026-01-31

Bug Fixes

• Install ko and crane from binary releases by @mchmarny

[0.1.2] - 2026-01-31

Bug Fixes

• Remove KO_DOCKER_REPO that conflicts with goreleaser repositories by @mchmarny

Other

• Restore flat namespace for container images by @mchmarny

Refactor

• Extract E2E tests into reusable composite action by @mchmarny

[0.1.1] - 2026-01-31

Bug Fixes

• Ko uppercase repository error and refactor on-tag workflow by @mchmarny

Refactor

• Migrate container images to project-specific registry path by @mchmarny

[0.1.0] - 2026-01-31

Bug Fixes

• Correct serviceAccountName field casing in Job specs by @mchmarny
• Add actions:read permission for CodeQL telemetry by @mchmarny
• Add explicit slug to Codecov action by @mchmarny
• Make SARIF upload graceful when code scanning unavailable by @mchmarny
• Install ko from binary release instead of go install by @mchmarny
• Strip v prefix from ko version for URL construction by @mchmarny

CI/CD

• Run test and e2e jobs concurrently by @mchmarny
• Add notice when SARIF upload is skipped by @mchmarny

Features

• Replace Codecov with GitHub-native coverage tracking by @mchmarny

Refactor

• Integrate E2E tests into main CI workflow by @mchmarny
• Split CI into unit, integration, and e2e jobs by @mchmarny

Tasks

• Init repo by @mchmarny
• Replace file-existence-action with hashFiles by @mchmarny
• Replace ko-build/setup-ko with go install by @mchmarny
• Remove Homebrew and update org to NVIDIA by @mchmarny
• Update settings by @mchmarny
• Remove code owners for now by @mchmarny
• Update project docs and setup by @mchmarny
• Update contributing doc by @mchmarny
• Remove badges not supported in local repos by @mchmarny