afgo-dsh: Integrating the AsterFire Go scientific computing platform into DSH as a workflow plugin

afgo-dsh is a DSH plugin maintained by YueSunhao. It is a pure Python standard library client (Python 3.8+, MIT license, 0 stars) that directly calls the Galileo Open API v1 of the Starlight Computing Platform AsterFire Go (sidereus-ai.com) (/open/v1, Asterfire-Token header), without needing to download the official asterfire Go binary or run `pip install`. It is released in two forms: as a DSH plugin (bundle) and as a general Agent skill. After installing the plugin and restarting DSH, five new tools—afg_platform, afg_files, afg_tasks, afg_workflows, and afg_raw—are added to the tool list. In skill form, it can be installed by cloning to the skill directory of each Agent. Its capabilities cover account/model queries, 71 Agents and 393 Kit directories, cloud drive upload/download, Agent task and Kit submission tracking, full workflow operations, and a raw arbitrary endpoint fallback. The README lists a total of 74 endpoints. Authentication requires only a token request header. The token is recommended to be placed in `~/.afg/config.json`; do not write it into the `token` field of `cordis.patch.yml`. The installation command is `dsh plugin --profile desktop add github:YueSunhao/afgo-dsh`. Please note: Chunked upload for large files is currently not supported; local directory installation may cause `ERR_MODULE_NOT_FOUND` due to junctions and should be changed to use a tarball or github/npm path. The plugin runs with the permissions of the current DSH process, so please read the source code and license carefully before installing.