feat(llm): add Atlas Cloud router

[binyangzhu000-sudo]

Description

Add Atlas Cloud as an opt-in OpenAI-compatible router for the direct-llm-agent baseline. The atlascloud/ prefix strips cleanly before requests, uses the hosted API URL by default, and reads credentials from ATLASCLOUD_API_KEY.

Atlas Cloud requests explicitly configure the OpenAI SDK with max_retries=0, preventing an implicit duplicate generation request. Existing TokenRouter, LiteLLM, and runner defaults are unchanged.

Type of Change

• New Benchmark Scenario (Industry/Asset type)
• Evaluation Metric / Scorer
• Agentic Orchestration Logic (ReAct, Plan-Execute, etc.)
• Infrastructure / Tooling Improvement

Industry Relevance

This adds another model endpoint for reproducible direct-model benchmark baselines while retaining the existing provider-prefixed model ID convention.

Related Issues

• Refs: N/A

Testing & Validation

• Unit Tests: focused src/llm/tests suite passed (25 tests).
• Scenario Validation: not applicable; no scenario definitions or trajectories changed.
• Data Integrity: no PII or industrial data added.

Additional validation: Ruff format/lint, mypy, py_compile, git diff --check, secret scan, current Atlas model catalog lookup, and one non-retried live chat-completions smoke request.

Checklist

• My code follows the project's Ruff formatting and linting rules.
• I have performed a self-review of my code.
• I have updated the documentation (README or /docs) accordingly.
• I have signed off my commits (DCO).

[DhavalRepo18]

@binyangzhu000-sudo Thanks for adding a support for atlas cloud. We added other router support as we get funding to run the model the respective platform host.

Do you pla to provide support ($$$) for Atlas cloud?