Agent Skills
› opensquilla/opensquilla
› meta-pdf-reformat-pipeline
meta-pdf-reformat-pipeline
GitHub将遗留PDF（如合同、手册）现代化，通过提取结构、重写问题页面、生成审计摘要并重新合并，实现文档内容的优化与格式更新。
Trigger Scenarios
需要更新老旧PDF内容
扫描版或法律文档的文本重构
PDF结构提取与重写
Install
npx skills add opensquilla/opensquilla --skill meta-pdf-reformat-pipeline -g -y
SKILL.md
Frontmatter
{
"kind": "meta",
"name": "meta-pdf-reformat-pipeline",
"always": false,
"triggers": [
"pdf 重排",
"pdf reformat",
"pdf 现代化",
"rewrite pdf"
],
"provenance": {
"origin": "opensquilla-original",
"license": "Apache-2.0"
},
"composition": {
"steps": [
{
"id": "extract",
"with": {
"task": "Extract structured text + page metadata from the PDF referenced in: {{ inputs.user_message | xml_escape | truncate(512) }}"
},
"skill": "pdf-toolkit"
},
{
"id": "rewrite",
"with": {
"instruction": "Modernize phrasing and standardize headings on problematic pages while preserving original meaning.",
"source_text": "{{ outputs.extract }}"
},
"skill": "nano-pdf",
"depends_on": [
"extract"
]
},
{
"id": "audit",
"with": {
"text": "{{ outputs.rewrite }}",
"style": "change_summary",
"max_words": 600
},
"skill": "summarize",
"depends_on": [
"rewrite"
]
},
{
"id": "merge",
"with": {
"task": "Merge the rewritten pages back into the source PDF. Audit summary for review: {{ outputs.audit }}"
},
"skill": "pdf-toolkit",
"depends_on": [
"rewrite",
"audit"
]
}
]
},
"description": "Modernize a legacy PDF: structural extraction → natural-language rewrite of problem pages → audit summary → re-merge into the final PDF.",
"meta_priority": 45
}
PDF Reformat Pipeline (Meta-Skill)
Historical-contract / scanned-manual / legal-document modernization in
4 steps: extract → rewrite → audit → re-merge. The audit step's output
gives a human reviewer a diff-summary before the merge is finalized.
Fallback
Run pdf-toolkit extract → nano-pdf rewrite → summarize → pdf-toolkit merge manually.
Version History
- 7f72a32 Current 2026-07-05 18:41
