Add OrcaRouter as a named AI provider

[XiaoHuo888-hue]

Scope

Adds a dedicated OrcaRouter provider rather than relying on the generic OpenAI-compatible endpoint, mirroring how this repo already treats OpenRouter-style aggregators. Named routers such as orcarouter/auto pick an upstream per request. OrcaRouter is an OpenAI-compatible gateway that exposes 150+ models behind one API key, and it also provides gateway-level security controls for AI agents.

The provider is wired end-to-end following the existing Z.AI/OpenRouter pattern:

• orcarouter service constant + model prefix (orcarouter@orcarouter/auto)
• DEFAULT_ORCAROUTER_CONFIG (https://api.orcarouter.ai)
• API-key, URL, and default settings in Config.ts (credential-backed, no plaintext-only storage)
• New OrcaRouterAdapter — OpenAI-compatible /v1 path with model discovery from /v1/models
• PROVIDER_DEFINITIONS entry with an ownedSecretId (secret-storage migration included automatically)
• Schema-driven settings UI rows + numeric range, provider detection in ProviderHelpers

I'm an engineer on the OrcaRouter team.

Compatibility and privacy

• Existing settings/frontmatter keys remain compatible — new keys merge through DEFAULT_SETTINGS with safe empty defaults; no migration needed.
• Provider/model routing behavior is covered — orcarouter@orcarouter/auto resolves to the new adapter; registry, URL-suffix, and credential tests updated.
• Tool approval and result-selection flows are preserved — no tool-calling paths touched.
• No API keys, prompts, note contents, generated bundles, or local files are included.

Validation

• npm run format:check
• npm run lint
• npm run typecheck
• npm test -- --runInBand (193 passed)
• npm run build

Manual checks

• Model discovery: https://api.orcarouter.ai/v1/models returns 192 models including orcarouter/auto (verified live).
• Desktop/mobile Obsidian interaction not exercised in CI; local build, typecheck, lint, and unit suite are green.