How do I install the PDF Processing skill?

Run npx skillmds@latest add haifai-ai/pdf-processing in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is PDF Processing safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with PDF Processing?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is PDF Processing free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published PDF Processing?

Haifai-AI (@haifai-ai) published this skill. Their other Agent Skills are listed on their SkillMD profile.

PDF Processing by haifai-ai — Agent Skill

Name: PDF Processing
Author: Haifai-AI

PDF Processing

Create PDFs by writing Python (reportlab) and running it — designed fresh per document, same visual language as our decks and docs; or manipulate existing PDFs with pypdf/pdfplumber (merge, split, fill forms, extract). Use for any PDF request or PDF-based analysis task.

Haifai-AI d695686 1.9 KB Updated Sep 17, 2026

File contents

PDF processing

Baby Whale builds and manipulates real PDFs through Python code — write the script, run it with bash using $DSH_OFFICE_PYTHON.

Creating PDFs (reportlab)

Prefer platypus flowables: SimpleDocTemplate, Paragraph, Table/TableStyle, Spacer, PageBreak.
Design per document, no default theme: pick a dominant + accent for THIS topic (reuse a companion deck's constants when one exists). Define colors.HexColor constants once; body text stays near-black.
Type scale: title 24–28pt bold, H2 14pt bold in the dominant color, body Helvetica 10.5pt with 14pt leading, generous spacing before headings.
Tables quiet by default: header band in the dominant color (white bold), alternating 3% tint rows, LINEBELOW hairlines only — never full grids.
Callout boxes: tinted single-cell tables, no border stripes.
Multi-page documents get page numbers + a small running footer via onPage callbacks.

Working with existing PDFs (pypdf / pdfplumber)

Merge/split/reorder pages with pypdf's PdfWriter.
Extract text with pdfplumber (page.extract_text()); extract tables with page.extract_tables(). For scanned pages without a text layer say so — do not invent contents.

Workflow

For from-scratch documents draft the spine first (see document-report), then encode it in the script.
Run via bash under $DSH_OFFICE_PYTHON; iterate on errors.
Verify: page count for writes, non-empty extracted text/pages for manipulations; keep console output to a couple of lines.
Save deliverables/<slug>.pdf, call deliver, state the path and a one-line summary of contents.

---name: pdf-processingdescription: Create PDFs by writing Python (reportlab) and running it — designed fresh per document, same visual language as our decks and docs; or manipulate existing PDFs with pypdf/pdfplumber (merge, split, fill forms, extract). Use for any PDF request or PDF-based analysis task.---# PDF processingBaby Whale builds and manipulates real PDFs through Python code — write thescript, run it with bash using `$DSH_OFFICE_PYTHON`.## Creating PDFs (reportlab)- Prefer `platypus` flowables: `SimpleDocTemplate`, `Paragraph`, `Table`/`TableStyle`, `Spacer`, `PageBreak`.- **Design per document, no default theme**: pick a dominant + accent for THIS topic (reuse a companion deck's constants when one exists). Define `colors.HexColor` constants once; body text stays near-black.- Type scale: title 24–28pt bold, H2 14pt bold in the dominant color, body Helvetica 10.5pt with 14pt leading, generous spacing before headings.- Tables quiet by default: header band in the dominant color (white bold), alternating 3% tint rows, `LINEBELOW` hairlines only — never full grids.- Callout boxes: tinted single-cell tables, no border stripes.- Multi-page documents get page numbers + a small running footer via `onPage` callbacks.## Working with existing PDFs (pypdf / pdfplumber)- Merge/split/reorder pages with pypdf's `PdfWriter`.- Extract text with pdfplumber (`page.extract_text()`); extract tables with `page.extract_tables()`. For scanned pages without a text layer say so — do not invent contents.## Workflow1. For from-scratch documents draft the spine first (see document-report), then encode it in the script.2. Run via bash under `$DSH_OFFICE_PYTHON`; iterate on errors.3. **Verify**: page count for writes, non-empty extracted text/pages for manipulations; keep console output to a couple of lines.4. Save `deliverables/<slug>.pdf`, call **deliver**, state the path and a one-line summary of contents.

Frequently asked questions