After running opc results config set, the entered configuration is stored as "extension" information in the ~/.kube/config. The configuration looks like this:

    extensions:
    - extension:
        api-path: ""
        apiVersion: results.tekton.dev/v1alpha2
        host: tekton-results-tekton-results.apps.<something.something>.openshiftapps.com
        insecure-skip-tls-verify: "false"
        kind: Client
        token: sha256~<token-value>
      name: tekton-results

When working with a different context (typically when using multiple tenants), that configuration has to be duplicated for each context. Is there a reason why this information is not associate with "cluster" rather than "context"? The token value is tied to specific context by being for a specific user used in the given context. However, the token value also just seems to duplicate what's already stored in the ~/.kube/config's users section. It seems making ocp read token form the users section (as it probably already does during the opc results config set step) would make the extension configuration context-independent and hence potentially movable to clusters configuration, avoiding the need to duplicate configuration for each context.

Additionally, there is this SRVKP-9156 that's causing the extension configuration to get lost on tenant change.