GPT Image Qwen Code 0.1.0-rc.2

RC2 fixes the independent GPT Image entry: ordinary native installation now starts the existing published mcp-openai-pro==2026.10.19.5 --profile gpt-image stdio package, with exactly five gpt_image_* tools and native sensitive-setting env injection. It does not connect to the general OpenAI remote or expose retired Sora tools.

Requires uv/uvx, Python >=3.10 and Node >=22. Both language guides cover native hidden input, current permissions/pricing, and same-task retrieval. Wait for a verified official Qwen release containing #14544 before entering a real Token; unfixed official0.25.2 remains synthetic-only. This is a Git source prerelease, not runtime GA or an official marketplace listing.

Validation: 4/4 Python unittest discovery checks and repository CI pass; existing published package synthetic stdio initialize/tools/list returns exactly five schemas. Actual new-tag native installation will be reported separately after execution-fixture fail-closed cleanup verification. No model, tools/call, media generation or billing validation. Old v0.1.0-rc.1 and the other four services are unchanged.

0.1.0-rc.1 — native Qwen source prerelease

Native Qwen Code source prerelease0.1.0-rc.1.

Install this independent service with qwen extensions install AceDataCloud/GPTImageQwenCode --ref v0.1.0-rc.1, or use the repository's defaultmain. The manifest, English/Chinese onboarding and dedicated sensitive setting are included.

Runtimegate: officialQwen0.25.2 does not load stored extension secrets into MCPHeaders. Until an officialfixedrelease of QwenLM/qwen-code#14544 is verified, use only a syntheticplaceholder for installation review; do not enter realTokens or call hostedmedia. This is a sourceprerelease, not authenticated-mediaGA or an officialQwenlisting. No production/model/paidgeneration was performed.

源码预发行：独立服务安装、双语图文和敏感凭据设置已包含；官方鉴权修复尚未发布，当前只用合成占位值验安装，不用真实Token或执行托管媒体。不是Qwen官方收录或媒体正式可用声明。