Support connector-scoped OpenAI-compatible request behavior

[dfliess]

Summary

• add connector-scoped structured_output_mode with the existing json_schema behavior as the default and a json_object fallback that includes the schema in the prompt
• add a validated extra_body map for provider-specific chat completion fields while rejecting overrides of core request and response-shape fields
• canonicalize and hash connection cache configuration so nested JSON values cannot alias and credentials are not embedded in cache keys
• document configuration for OpenAI-compatible endpoints

Motivation

OpenAI-compatible providers differ in structured-output support and provider-specific request extensions. This was motivated by an interoperability report in kairosagentica/kairos-cloud#145.

The defaults remain unchanged for existing OpenAI connectors, and all new behavior is configured per connector.

Test plan

• go test -race ./runtime/drivers/openai -run "TestComplete(AppliesConnectorRequestBehavior|PassesNestedExtraBody|DefaultsToJSONSchema)|TestOpenValidatesProviderRequestBehavior" -count=1
• go test ./runtime -run TestGenerateConnectionKey -count=1
• go test ./runtime/... -run "^$"
• go vet ./runtime/drivers/openai ./runtime
• git diff --check upstream/main...HEAD