Aidenwu0209/dsh-PaddleOCR-Skills
PaddleOCR skills for DeepSeek Harness with native tools and GUI configuration
A native DeepSeek Harness bundle with two native tools — paddleocr_text_recognition (image/PDF text extraction) and paddleocr_doc_parsing (layout, tables, Markdown, document structure) — plus two skills and a dedicated Settings → PaddleOCR GUI. GUI fields for both endpoints, timeouts, uv, result storage, and a DSH Credential reference; tokens stored through DSH Credentials and never returned to the browser; real-path workspace containment; auditable raw JSON results under .dsh-paddleocr/results/. Adapted from PaddleOCR-Skills (4.3K+ installs across supported AI agents).
Install
npx @deepseek-ai/dsh plugin --profile web add "github:Aidenwu0209/dsh-PaddleOCR-Skills#main"README documents the one-prompt agent install (a terminal-capable AI agent runs the full flow) and the direct command npx @deepseek-ai/dsh plugin --profile web add "github:Aidenwu0209/dsh-PaddleOCR-Skills#main" — note the README prompt text drops the # in one spot; the code block uses github:Aidenwu0209/dsh-PaddleOCR-Skills#main. Requirements: Node.js 22.19+, DeepSeek Harness, Python 3.9+, and uv. Built lib/ artifacts are included. From OpenClaw: openclaw skills install @aidenwu0209/dsh-paddleocr-skills-setup installs the repository-specific setup guide.
Compatibility
DeepSeek Harness Web profile; Node.js 22.19+, Python 3.9+, uv. Two native tools + two skills + a Settings → PaddleOCR GUI.
Details
- Repo: Aidenwu0209/dsh-PaddleOCR-Skills
- Category: Coding & Development
- Stars: 4
- Version: GitHub source (github:Aidenwu0209/dsh-PaddleOCR-Skills#main; no npm package)
- Last push: 2026-09-02
- First seen: 2026-08-14
Recent updates
Adapted from PaddleOCR-Skills; native tools + skills + GUI; credential-safe token storage; auditable JSON results.
FAQ
- What are the prerequisites?
- Node.js 22.19+, DeepSeek Harness, Python 3.9+, and uv. The README's one-prompt agent install checks these first and uses official installers for anything missing.
- How are API tokens handled?
- Tokens are stored through DSH Credentials and never returned to the browser. The GUI shows clickable links to the PaddleOCR official site, API-token page, and official API documentation.
- What tools does it add?
- paddleocr_text_recognition for image/PDF text extraction and paddleocr_doc_parsing for layout, tables, Markdown, and document structure, plus two skills.
Alternatives
omdsh-dev/dsh-paddle-ocr · Aidenwu0209/dsh-Unlimited-OCR-Skill · BYYY-eng/deepseek-harness-file-upload-ocr-plugin