AWS Cloud Controller Manager never deploys after cluster creation - AdditionalManifest written with invalid format

[yonasberhe23]

Setup

• Rancher version: v2.15-8be6998d2bbf1225b9d74e495218e4e8ca15c125-head / Dashboard | master d3c8d5a
• Rancher UI Extensions: Capi Capa UI 0.0.0-rc6
• Browser type & version:

Describe the bug
When creating a CAPA cluster via the UI, the cluster gets stuck in  Updating  — "waiting for viable init node". The  additionalManifest  field is written with a duplicate  additionalManifest: |-  prefix, producing invalid YAML in  addons.yaml. As a result, the CCM HelmChart is never created and nodes remain tainted with  node.cloudprovider.kubernetes.io/uninitialized 

To Reproduce

1. Set up Rancher with CAPA UI extension 0.0.0-rc6 from https://rancher.github.io/prov-capi-ui-extensions
2. Create a CAPIProvider for AWS
3. Cluster Management → Create → AWS (CAPA) → fill in minimal fields (region, instance type, SSH key) → Save
4. Observe cluster stuck in  Updating  — "waiting for viable init node"
5. SSH into the control plane node:
   cat /var/lib/rancher/rke2/server/manifests/rancher/addons.yaml
6. Observe  addons.yaml  starts with  additionalManifest: |-  instead of a valid Kubernetes manifest
7. Confirm  aws-cloud-controller-manager  is missing from  kube-syste

Result
additionalManifest  value is double-wrapped, resulting in:

additionalManifest: |-
  additionalManifest: |-
        apiVersion: helm.cattle.io/v1
        kind: HelmChart
        ...

Cluster is stuck in  Updating  — "waiting for viable init node". Nodes remain tainted with  node.cloudprovider.kubernetes.io/uninitialized  and CCM is never deployed.

Expected Result
 addons.yaml  contains a valid HelmChart manifest and CCM is deployed automatically:

apiVersion: helm.cattle.io/v1
kind: HelmChart
metadata:
  name: aws-cloud-controller-manager
  namespace: kube-system

Screenshots

Screen.Recording.2026-07-15.at.6.57.56.PM.mov

Additional context
Related to #16540

Workaround
Fixed by removing the duplicate prefix