Set up DSH PaddleOCR Skills
Install the native DeepSeek Harness bundle. Do not present it as an OpenClaw code plugin; this setup skill guides installation into DSH.
Use this setup skill when the request mentions DeepSeek Harness, DSH,
the Settings → PaddleOCR GUI, or tasks such as 图片转文字 / 截图识字 /
扫描件 OCR / PDF 转 Markdown / 表格提取 / 公式识别 / 版面分析. The bundle
provides native tools for plain OCR and structured document parsing, plus GUI
fields for service endpoints, credentials, timeouts, uv, and result storage.
Install and launch
-
Check Node.js 22.19+, Python 3.9+,
npx, anduv. Explain any missing prerequisite before using its official installer. Do not usesudoor change unrelated settings without permission. -
Run:
npx @deepseek-ai/dsh plugin --profile web add "github:Aidenwu0209/dsh-PaddleOCR-Skills#main" npx @deepseek-ai/dsh web -
Wait for the real local Web URL, open it, and confirm that Settings → PaddleOCR is visible.
-
Confirm the settings panel links to https://www.paddleocr.com, the API-token page, and official API documentation.
Configure safely
- Ask the user for the HTTPS
/ocrand/layout-parsingendpoints they need. - Store the token through the DSH Credential field; never echo, log, or place it in ordinary settings or source files.
- Confirm
uv, the selected endpoint, and credential status in the GUI. - Do not claim success unless the plugin install succeeded, the Web URL responded, and the Settings panel was actually visible.
Report prerequisite versions, commands, the Web URL, visible configuration status, and any values still required from the user.
