Add Atlas Cloud LLM provider

[binyangzhu000-sudo]

Summary

• add Atlas Cloud as a built-in OpenAI-compatible LLM provider
• default Atlas Cloud to https://api.atlascloud.ai/v1 and qwen/qwen3.5-flash, with deepseek-ai/deepseek-v4-pro in the fallback model list
• support ATLASCLOUD_API_KEY plus ATLAS_CLOUD_API_KEY aliases for API key, base URL, and model env fallback
• keep Atlas Cloud out of the RAG embedding provider list because this integration targets chat/completions LLM usage only

Tests

• corepack pnpm install --frozen-lockfile --ignore-scripts
• corepack pnpm --filter @ai-novel/server prisma:generate
• corepack pnpm --filter @ai-novel/shared build
• corepack pnpm --filter @ai-novel/server build
• node --test server/tests/llmProviders.test.js
• corepack pnpm --filter @ai-novel/server lint
• corepack pnpm --filter @ai-novel/client typecheck
• git diff --check
• git diff --cached --check

Note: the focused Node test logs the existing local SQLite APIKey table-missing message while exercising the no-database fallback path, but all assertions pass. Atlas Cloud live model catalog was checked and contains both qwen/qwen3.5-flash and deepseek-ai/deepseek-v4-pro.