Skip to main content

Release 0.5.11 — Setup Wizard, Ollama + LiteLLM, and RBAC Fixes

Sri Aradhyula
Sri Aradhyula

Released: 2026-06-11 Chart: oci://ghcr.io/cnoe-io/charts/ai-platform-engineering:0.5.11 Previous release: 0.5.10

Highlights​

0.5.11 makes the self-hosted setup experience significantly smoother: the setup script now auto-installs missing tools so you no longer need to prepare your environment manually, and Ollama users can now configure a LiteLLM proxy during setup. This release also fixes a set of access-control edge cases that could prevent organization members from discovering skills or reaching RAG endpoints.

What's New​

Setup Experience​

  • Automatic prerequisite installation — setup-caipe.sh detects missing tools (kind, kubectl, helm, etc.) and installs them for you, gating any sudo steps on your explicit consent rather than failing silently
  • LiteLLM proxy support for Ollama — when you choose Ollama as your model backend, the setup wizard now asks whether you want to route requests through a LiteLLM proxy, giving you OpenAI-compatible routing without extra manual steps

Access Control​

  • AD group admins as platform super-admins — if your organization already has Active Directory group admins, you can now map them directly to the platform super-admin role without duplicating team membership

Bug Fixes​

  • Skills not discoverable by org members — organization members could not see skills listed in their workspace; all org members are now included in the reader and user type restrictions for skills
  • RAG endpoints blocked for org-admins — org-admin users were incorrectly denied access to admin-gated RAG endpoints; they now receive the correct elevated role
  • Ollama setup: fixed FQDN resolution, RAG ingestor client wiring, and bearer auth for dynamic agents when Ollama is selected
  • RAG server: OIDC provider credentials and ingestor client settings are now correctly passed through to the RAG server on deployment

Breaking Changes​

No breaking changes. Drop-in upgrade from 0.5.10.

Upgrade​

helm upgrade ai-platform-engineering \
  oci://ghcr.io/cnoe-io/charts/ai-platform-engineering \
  --version 0.5.11 \
  -f your-values.yaml

No Helm values changes between 0.5.10 and 0.5.11. No data migrations required.