Read, fill, extract, and generate PDF documents from an agent session — forms, reports, and print-ready files rather than markdown-only output.
Get the complete skill folder, including its scripts and reference files.
·
Turn a PDF table into a usable Excel sheet
Extract six invoice lines, keep their source references, and reconcile the result before using it. (2 sample files)
When to use this skill
Use the PDF skill when the job is a real PDF: fill a form, extract tables, stamp a page, or produce a print-ready report. It is the right skill for "turn this into a PDF" and "what does this PDF say".
When not to use it
Use DOCX when the collaborator needs an editable Word file, XLSX for spreadsheets, and PPTX for slides. Do not use this skill to generate long prose that should stay in markdown.
How to install
Where will you use it?
Install the publisher’s plugin
This installs a bundle that can include other skills. Choose this method or the manual alternative below; you only need one installation.
Run these commands in Claude Code:
Claude Code CLI/plugin marketplace add anthropics/skills /plugin install document-skills@anthropic-agent-skills
- Ask Claude Code to use Pdf on a small task and confirm that it loads. Check the setup requirements in the publisher’s guide if it cannot run.
Alternative: install the folder manually
Use these steps if you choose the manual folder method. If you installed the plugin above, skip this alternative.
- Claude Code personal install: place the skill and its bundled files at ~/.claude/skills/pdf/SKILL.md. For a shared project, commit the folder at .claude/skills/pdf/ instead.
- Current Claude Code detects SKILL.md changes in watched directories during the session. If you just created a top-level skills directory, run /reload-skills. Check that the skill is listed before testing its trigger.
- For Claude chat or Cowork in the Desktop app, manage uploaded skills in Customize > Skills; copying files into a local Claude Code folder is not a Desktop install. Other agents may use different install locations.
What supports this recommendation?
75/100 provisional content score. This measures documentation, practical detail and source signals. It is not a measured task success rate or a security certification.
- Documentation
- 77/100
- Practical detail
- 68/100
- Source evidence
- 70/100
- Source signals
- 92/100
Model compatibility evidence
- claudeuntested
No model-specific signal or recorded compatibility test was found.
- chatgptuntested
No model-specific signal or recorded compatibility test was found.
- geminiuntested
No model-specific signal or recorded compatibility test was found.
- copilotuntested
No model-specific signal or recorded compatibility test was found.
- llamauntested
No model-specific signal or recorded compatibility test was found.
- perplexityuntested
No model-specific signal or recorded compatibility test was found.
- mistraluntested
No model-specific signal or recorded compatibility test was found.
- grokuntested
No model-specific signal or recorded compatibility test was found.
Inferred compatibility is based on content signals. Check the app, its available tools and setup requirements before running a skill.
Try one small task
Best first exercise: extract a small text-based table and reconcile every row against the PDF.
See actual outputs and execution limits →What you need
- An agent session with local file access and a Python runtime; the upstream guide uses pypdf and pdfplumber.
- The full PDF skill folder, including its reference files. Use the upstream installation instructions for your agent.
- Download the synthetic invoice PDF from the PDF-to-Excel workflow below.
Get the sample input and full checklist
Extract six invoice lines without changing IDs, quantities, or currency.
Use the PDF skill to read invoice-lines.pdf. Extract the six line items into CSV with columns item_id, description, quantity, unit_price_usd, line_total_usd, source_page. Preserve the IDs including leading zeros. Reconcile each line and the printed subtotal. Report unreadable values instead of guessing. Return the CSV file and a short reconciliation note.
Expected result
A specification to check against, not output from an AI run.
A six-row CSV whose line amounts total USD 1,250.00, plus a note that all rows came from page 1.
- Exactly six items, 001 through 006, with source_page equal to 1.
- Quantities and unit prices remain numeric; item IDs preserve their leading zeros.
- The extracted line totals sum to the PDF subtotal of USD 1,250.00.
Where this exercise stops
- This fixture contains selectable text. Passing it says nothing about OCR accuracy on scans.
- Encrypted PDFs, unusual layouts, and forms require different handling in the upstream guide.
Recorded example run
These files came from actual execution in a Codex session. They were produced and checked within the same session; this is scoped evidence, not independent review or a measured success rate.
PDF invoice extracted and reconciled in Excel
Executed · Automated Codex exercise
Codex agent in this session; Python 3.12.3, openpyxl 3.1.5, LibreOffice 24.2.7.2 420(Build:2); aarch64. PDF extraction: pdfplumber 0.11.10.
The agent read the pinned upstream instructions and applicable helpers, then authored and executed the committed scripts locally. This is not a native Claude skill-loader test or an independent benchmark.
Download the actual outputs
Recorded checks and observations
- Passed: All six extracted rows match the CSV answer key; leading-zero IDs are preserved.
- Passed: The recalculated subtotal is USD 1,250.00 with a USD 0.00 reconciliation difference.
- Passed: Nine formulas recalculate without errors. Changing a quantity in a copy updates the subtotal to USD 1,295.00.
Scope, limitations and pinned sources
- The agent read the pinned upstream instructions and used their prescribed tools. Native Claude skill loading and general cross-agent compatibility were not tested.
- One small synthetic input per workflow; results are not a success rate or a broad model-compatibility claim.
- The Codex agent that produced the outputs also ran and interpreted the checks; no independent human review occurred.
- PDF extraction used selectable text and regular table borders. OCR, merged cells, and multi-page PDFs were not tested.
- Spreadsheet recalculation used LibreOffice 24.2.7.2; Microsoft Excel was not run.
- This session already knew the published acceptance criteria. The CSV answer key was read by verification after PDF extraction, but this is not a blind or held-out evaluation.
Source and evidence
The upstream instructions informed this guide. The exercise and acceptance checks are an OpxSkills test recipe. Where a run is recorded, its actual outputs, environment and limits appear separately above. A recorded example does not establish general model compatibility or a success rate.
Source reviewed · Upstream revision 8a1541c4a3ff
What this skill does
Claude's document skills treat PDFs as a first-class artifact. The PDF skill covers extraction, form fill, and generation with the same progressive-disclosure pattern as other SKILL.md packages.
Pair it with a project-level skill if your team has a house cover page or filename convention.
Intended uses
- Fill a vendor PDF form from structured notes
- Extract tables from a scanned report
- Export an agent-written brief as a print-ready PDF
Implementation notes
- Say whether you need a filled form or a new file.
- Provide the source PDF in the session; do not ask the model to guess fields.
- For Excel-like tables, confirm whether XLSX is the better output.
Try it, then tell us how it went
A copied prompt does not tell us whether the task worked. Report your result after checking the output. We collect your tool and outcome, never your files or prompt.