from openspace5,230
Generate DOCX/PDF/HTML from Markdown using Pandoc by default; fallback to ReportLab for heavy Unicode or programmatic layouts. Includes sanitization and backend
This skill provides a robust document-generation workflow that prefers direct Pandoc execution for standard Markdown→DOCX/PDF/HTML conversions and uses ReportLab as a reliable fallback for heavy Unicode or programmatic PDF needs. It outlines quick-start steps, Unicode sanitization, two backend paths (Pandoc and ReportLab), troubleshooting tips, and sample scripts for typical conversion tasks.
Use the Pandoc path for most straightforward conversions (DOCX, HTML, PDF when LaTeX is available). Choose the ReportLab path when documents contain heavy Unicode (CJK, special symbols), require precise programmatic layout, or when LaTeX/pandoc dependencies are unavailable. Use the dual-backend workflow when you need automated fallback and error recovery between backends.
Suitable for agents that can run shell commands and write files (shell-enabled agents such as Claude Code, Copilot/Codex-like environments, Cursor/Gemini CLI integrations).
Document generation skill using pandoc as default backend with ReportLab fallback for Unicode-heavy content. No bundled scripts — purely instruction-based SKILL.md. Well-documented with decision flowchart and backend selection table, but suffers from content duplication and padding. The sed sanitization snippet has minor shell injection risk with unquoted variables, but no executable scripts are bundled.
Redundant content: Examples 0-2 and troubleshooting sections appear duplicated in different forms. SKILL.md could be 40% shorter without losing value. The dual-backend approach with Unicode handling guidance is genuinely useful and well-thought-out. Source is from an academic benchmark repo (hkuds/openspace gdpval_bench) which may explain the padding.
