feat: add EvoLink OpenAI-compatible provider

[EvoLinkAI]

TLDR

Adds EvoLink as a first-class OpenAI-compatible provider for the main OpenGame agent LLM. EvoLink uses the existing OpenAI-compatible flow with OPENAI_BASE_URL=https://direct.evolink.ai/v1, so the provider inherits the default request/client behavior and only needs provider detection, registration, tests, and configuration docs.

Dive Deeper

• Adds EvoLinkOpenAICompatibleProvider, selected when the configured base URL contains direct.evolink.ai.
• Registers and exports the provider through the existing OpenAI-compatible provider factory.
• Adds offline unit tests for provider detection and factory selection.
• Documents EvoLink setup in the README, quickstart, auth docs, provider README, and .env.example.

Reviewer Test Plan

Use an EvoLink API key with the OpenAI-compatible auth flow:

export OPENAI_API_KEY="your-evolink-api-key"
export OPENAI_BASE_URL="https://direct.evolink.ai/v1"
export OPENAI_MODEL="gpt-5.2"
opengame -p "Build a Snake clone with WASD controls and a dark theme." --yolo

The provider factory should select EvoLinkOpenAICompatibleProvider, while requests continue through the standard OpenAI-compatible chat completions path.

Testing Matrix

	🍏	🪟	🐧
npm run	✅	❓	❓
npx	❓	❓	❓
Docker	❓	❓	❓
Podman	❓	-	-
Seatbelt	❓	-	-

Validated locally on macOS with Node v24.13.1:

• npm run test --workspace=packages/core -- src/core/openaiContentGenerator/provider src/core/openaiContentGenerator/index.test.ts — 7 test files / 75 tests passed.
• npx eslint packages/core/src/core/openaiContentGenerator/index.ts packages/core/src/core/openaiContentGenerator/index.test.ts packages/core/src/core/openaiContentGenerator/provider/index.ts packages/core/src/core/openaiContentGenerator/provider/evolink.ts packages/core/src/core/openaiContentGenerator/provider/evolink.test.ts --max-warnings 0 — passed.
• npm run typecheck --workspace=packages/core — passed.
• npm run build --workspace=packages/core — passed.
• Live provider-path verification with gpt-5.2 returned provider=EvoLinkOpenAICompatibleProvider, model=gpt-5.2-2025-12-11, content=OK.

Linked issues / bugs

Resolves #44

[EvoLinkAI]

Validation evidence from local macOS run:

npm run test --workspace=packages/core -- src/core/openaiContentGenerator/provider src/core/openaiContentGenerator/index.test.ts
Test Files  7 passed (7)
Tests       75 passed (75)

npx eslint packages/core/src/core/openaiContentGenerator/index.ts packages/core/src/core/openaiContentGenerator/index.test.ts packages/core/src/core/openaiContentGenerator/provider/index.ts packages/core/src/core/openaiContentGenerator/provider/evolink.ts packages/core/src/core/openaiContentGenerator/provider/evolink.test.ts --max-warnings 0
passed with no output

npm run typecheck --workspace=packages/core
> tsc --noEmit
passed

npm run build --workspace=packages/core
> node ../../scripts/build_package.js
Successfully copied files.

Live EvoLink call through the provider selection path:

provider=EvoLinkOpenAICompatibleProvider
model=gpt-5.2-2025-12-11
content=OK