feat: Support additional create methods for agent and agent_graph

[jsonbailey]

fix!: Rename default_ai_provider parameter → default_connector on all LDAIClient methods
feat!: Rename AIProvider → AIConnector
feat!: Rename AIProviderFactory → RunnerFactory, create_model/agent/agent_graph()
feat!: Rename OpenAIProvider → OpenAIConnector
feat!: Rename LangChainProvider → LangChainConnector

---

Note

Medium Risk
Broad rename/refactor across the SDK and provider packages (interfaces, factories, and public method parameters) introduces breaking API changes and could impact runtime backend selection. Behavior is mostly preserved, but new connector instantiation defaults (e.g., OpenAI client from env) and fallback creation paths increase integration risk.

Overview
Refactors the SDK/provider integration from AIProvider to a new per-backend AIConnector abstraction. This renames AIProvider/AIProviderFactory to AIConnector/RunnerFactory, updates Chat/Judge/LDAIClient to depend on connectors, and changes the public override parameter from default_ai_provider to default_connector.

Adds a new RunnerFactory capability factory that can create model, agent, and agent_graph runners with a shared fallback mechanism across backends.

Updates backend packages to match the new connector model. OpenAIProvider/LangChainProvider become OpenAIConnector/LangChainConnector, adopting create_model() instance factories; OpenAI now lazily builds an AsyncOpenAI client from OPENAI_API_KEY when not provided. Tests are updated accordingly and legacy provider/factory files are removed.

^{Written by Cursor Bugbot for commit dd194a8. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}