Add optional Atlas Cloud image provider

[binyangzhu000-sudo]

Summary

• add an optional --provider atlascloud path for text-to-image generation through Atlas Cloud's async media API
• map the existing prompt, size, quality, format, and moderation flags to the live openai/gpt-image-2/text-to-image schema
• document the Atlas Cloud key and text-to-image-only scope in the CLI README, Chinese README, and skill metadata

Validation

• verified openai/gpt-image-2/text-to-image is present in the live Atlas Cloud model list with display_console=true
• verified the live schema supports model, prompt, size, quality, output_format, and moderation

Tests

• python3.11 -m py_compile src/gpt_image_cli/cli.py skills/gpt-image/scripts/generate.py
• PYTHONPATH=src python3.11 -m gpt_image_cli.cli --help
• env -u ATLASCLOUD_API_KEY -u ATLAS_CLOUD_API_KEY PYTHONPATH=src python3.11 -m gpt_image_cli.cli --provider atlascloud -p 'test prompt' -f /tmp/atlas-no-key.png