DosterBool/dsh-ocr-seam-guard
DosterBool★ 0JavaScriptLast synced: 2026-08-16
Self-healing guard for dsh-ocr: auto re-applies the llm-deepseek adapter OCR seam and plugin fixes wiped by dsh upgrades
README excerpt
dsh-ocr-seam-guard Self-healing guard for the dsh-ocr plugin on DeepSeek Harness: automatically re-applies the local patches that npm i -g @deepseek-ai/dsh upgrades (or a plugin re-install) wipe out. 中文说明见下方。 What it guards The dsh-ocr plugin needs three fixes that live OUTSIDE the profile (so profile reconcile never touches them, and a harness upgrade wipes the adapter one): Fix File Wiped by The ocr service seam in the official dsh-llm-deepseek adapter /node modules/@deepseek-ai/dsh-llm-deepseek/lib/index.js npm i -g @deepseek-ai/dsh upgrade UTF-8 stdout enforcement for the Python OCR subprocess /profiles/node modules/@deepseek-ai/dsh-ocr/lib/index.js plugin re-install rapidocr 1.2.3 API fix ( text det → text detector , text rec → text recognizer ) /profiles/node modules/@deepseek-ai/dsh-ocr/tools/ocr image.py plugin re-install How it works On every harness boot the plugin checks the three sites. A missing fix is re-applied with anchored string replacements — every anchor must match exactly once or the whole site is skipped (never a partial patch) — then the patched module is import-verified in a child node process. A failed verification rolls the file back and reports the error.…
View full README on GitHub →Category
The first vision plugin for DeepSeek Harness, and the vision bridge for every text-only coding agent. Paste an image, get structured JSON evidence (OCR, layout, semantics). | 全网最强 DeepSeek Harness 外挂视觉插件,为 DeepSeek、GLM 等纯文本模型外挂视觉能力,粘贴图片即得结构化 JSON 证据(OCR、版面、语义)。
★ 2,208
Anionex/agent-vision-toolkit为纯文本模型"看图“设计更好的视觉工具箱和技能,支持多图理解,图片问答,前端UI还原、GUI 自动化等,并可选无缝接入多个主流agent,直接识别粘贴图片| A vision toolkit and skill designed for text-only llms — image Q&A, long-screenshot OCR, frontend UI restoration, and GUI automation, with optional seamless integration for Codex, Claude Code, Pi, Oh My Pi, and OpenCode
★ 933
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.
★ 488