开源Agent技能scan-to-vector发布
Start here
2026年11月4日，面向Claude Code与Codex的开源Agent技能scan-to-vector正式发布。该工具能将扫描的书籍、手册与图表重构为包含矢量图形与可搜索文本的高清PDF。其实现原理是结合本地脚本与OCR技术，将线框和图表重绘为矢量路径，并与原始扫描件逐页叠加大纲进行审核。测试集结果显示，该技能使Haiku 5.5的长文档得分从0.841提升至0.997，同时使Sonnet 5.5的处理耗时减少了37%。
AI-generated from reports · Updated 3 hours ago
Report timeline
Follow the reports to see different sides of the story.
- Hacker News · AISelectedOpen-source agent skill scan-to-vector: Reconstructing scanned documents into vector PDFs
Designed for Claude Code and Codex, scan-to-vector is an agent skill that reconstructs scanned books, manuals, and diagrams into high-resolution PDFs featuring vector graphics and searchable text. Combining local scripts with OCR, the tool redraws line art and diagrams as vector paths and conducts page-by-page outline verification against the original scans. In test sets, the skill boosted Haiku 5.5's long-document score from 0.841 to 0.997 and cut Sonnet 5.5's execution time by 37%.
Story heat trend
Current heat 9·Comparable peak 10（Nov 4 23:00）·Comparable change, last 24 hours–
Trends compare only the same parties observed continuously, so the scope can be narrower than current heat counts. Move the pointer or tap the chart to see hourly heat; use the left and right arrow keys on a keyboard.