refactor: flatten providers config to use type as active provider selector

[avoidwork]

Summary

Flatten the providers config in config.yaml by moving the type field from providers.openai.type to providers.type, making it the active provider selector with provider-specific configs as sibling keys.

Motivation

The current structure nests type inside each provider config (e.g., providers.openai.type: openai), which is redundant and does not scale well when adding new providers. Moving type to the providers level makes it clear which provider is active, and provider-specific configs live as sibling keys keyed by their type name.

Changes

• config.yaml: Added type: openai at the providers level, removed type: openai from providers.openai
• src/config/schemas.js: Removed type from _OpenaiProviderConfigSchema, added type enum to ProvidersSchema

Proposed Structure

Before:

providers:
  openai:
    type: openai
    base_url: https://api.openai.com/v1
    model: gpt-4o

After:

providers:
  type: openai
  openai:
    base_url: https://api.openai.com/v1
    model: gpt-4o

Testing

• All 1209 tests pass
• Lint passes (oxlint + oxfmt)
• No code changes needed — existing dynamic access patterns (config?.providers?.[providerName]) are already compatible

Related

Closes #483