feat(llm): add Atlas Cloud environment fallback

[binyangzhu000-sudo]

Summary

• support ATLASCLOUD_* and ATLAS_CLOUD_* environment variables when the generic LLM_* configuration is absent
• default the Atlas shortcut to https://api.atlascloud.ai/v1 and deepseek-ai/deepseek-v4-pro
• preserve the existing priority order: explicit constructor arguments, then generic LLM_*, then Atlas-specific fallback
• add focused tests for defaults, aliases, and generic configuration precedence

Configuration

ATLASCLOUD_API_KEY=your-atlas-cloud-api-key
ATLASCLOUD_BASE_URL=https://api.atlascloud.ai/v1
ATLASCLOUD_MODEL=deepseek-ai/deepseek-v4-pro

The existing LLM_API_KEY, LLM_BASE_URL, and LLM_MODEL_ID configuration remains unchanged and takes priority.

Validation

• focused LLM tests: 10 passed, 3 skipped
• Python compileall and git diff --check passed
• live HelloAgentsLLM chat completion through Atlas Cloud: exact atlas-ok response, 46 total tokens
• live Function Calling through Atlas Cloud: one calculate tool call with { "expression": "2+3" }, 309 total tokens
• broader suite excluding the pre-existing collection blocker: 165 passed, 34 failed, 3 skipped, 4 errors; failures are existing environment/test issues such as missing pytest-asyncio, tests constructing an LLM without credentials, and unrelated tool assertions
• full collection is also blocked on the existing PlanAndSolveAgent import mismatch (PlanSolveAgent is exported)

No README, logo, sponsor, credits, or partner changes.