GitHub · 界面与桌面
Anionex/dsh-vision-toolkit
让纯文本模型更好地做视觉任务的DeepSeek Harness插件：带意图的图片问答、长截图 OCR、UI 还原等｜DeepSeek Harness-native integration for agent-vision-toolkit: image Q&A, long-screenshot OCR, UI restoration, grounding, pixel diff, Artifacts, and Web UI.
SOURCE-DECLARED INSTALL
dsh plugin --profile web add github:Anionex/dsh-vision-toolkitSource-declared command. Review the package and Registry Listing before running it.
MEDIA REFERENCES
Captured in public view
REGISTRY LISTINGS
Source-attributed install evidence
| Registry source | Install spec | Version | Verification claim | Observed | Raw snapshot |
|---|---|---|---|---|---|
| zoahdev/dsh-subscribe · registry.json | github:Anionex/dsh-vision-toolkitgit | 0.1.7 | source claims verified | 2026-08-15 → 2026-08-16 | 962 |
| awesome-dsh-plugin/awesome-dsh-plugin · https://awesome-dsh-plugin.com/plugins.json | @anionex/dsh-vision-toolkitnpm | — | unreported | 2026-08-15 → 2026-08-16 | 1093 |
| beancookie/awesome-dsh-plugin · docs/plugins.json | @dsh-external/dsh-vision-toolkitnpm | — | unreported | 2026-08-15 → 2026-08-16 | 1093 |
| dsh-market/dsh-market · data/registry-snapshot.json | @anionex/dsh-vision-toolkitnpm | — | unreported | 2026-08-15 → 2026-08-16 | 962 |
| Zhiyuan-Fan/Awesome-DeepSeek-Harness-Plugins · README.md | — | — | unreported | 2026-08-15 → 2026-08-16 | 1093 |
| walkinglabs/awesome-deepseek-harness-plugins · README.md | — | — | unreported | 2026-08-15 → 2026-08-16 | 1093 |
| AdamPlatin123/awesome-dsh-plugins · README.md | — | — | unreported | 2026-08-15 → 2026-08-16 | 962 |
Verification is a claim made by the named Registry Source; it is not a security, compatibility, quality, or official-endorsement claim by this site.
CONTEXT
Why it is here
让纯文本模型更好地做视觉任务的DeepSeek Harness插件：带意图的图片问答、长截图 OCR、UI 还原等｜DeepSeek Harness-native integration for agent-vision-toolkit: image Q&A, long-screenshot OCR, UI restoration, grounding, pixel diff, Artifacts, and Web UI.
Evidence updated 2026-08-16T13:42:34Z. Interaction numbers are platform-native snapshots; the evidence panel records the metric source and observation time. NULL means the public page did not expose a number at collection time.