An agent skill by anthropics, from anthropics/skills. Tags: anthropic, debugging, documents, office.
What it does
Use this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill.
Install
With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:
npx skills add anthropics/skills --skill pdfOr copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):
git clone --depth 1 https://github.com/anthropics/skills
cp -r skills/skills/pdf ~/.claude/skills/pdfSafety box score
Grade B, low risk: 18 of 100 risk points, rated 2026-09-20. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.
This skill processes PDFs locally: it can read/extract text, render pages to images, and create, merge, encrypt, decrypt, or fill PDFs. The main risks are unrestricted writes to user-chosen paths and optional OCR workflows that depend on third-party packages; it does not show built-in remote-host access, privilege escalation, or secret-store access.
| Category | Level | Why |
|---|---|---|
| Code & shell execution | 1 (low) | The skill ships runnable Python utilities and tells the agent to invoke them, e.g. `python scripts/fill_fillable_fields.py ...`, but it does not evaluate downloaded code or shell pipelines. |
| Network access | 1 (low) | The OCR path depends on external packages (`pip install pytesseract pdf2image`), so using that branch brings in network-fetched third-party code even though the shipped scripts themselves do not call remote hosts. |
| Filesystem reach | 2 (moderate) | Several scripts write PDFs, JSON, and PNGs to caller-supplied paths (for example `with open(output_pdf_path, "wb")` and `image.save(image_path)`), so an agent can overwrite any writable file path you point it at. |
| Secrets & credentials | 0 (none) | The files only read PDFs/JSON and do not reference keychains, browser sessions, CLI logins, or credential stores. |
| Privilege & persistence | 0 (none) | There is no `sudo`, global install, service registration, shell startup editing, or other privilege-changing step in the shipped files. |
| Instruction hijack surface | 0 (none) | The instructions are procedural (`"CRITICAL: You MUST complete these steps in order"`) but do not tell the agent to ignore higher-priority rules, hide actions, or follow remote instructions. |
| Irreversible actions | 1 (low) | The skill can automatically merge, fill, encrypt, decrypt, and annotate PDFs, changing document contents without a separate human approval gate, though that is inherent to the skill's purpose. |
| Transparency | 0 (none) | All shipped files are readable text; there are no binaries, minified blobs, or unpinned remote installers in the repository. |
Before installing
- Use a throwaway output directory first; the scripts write to the paths you give them.
- If you will use OCR, install `pytesseract`/`pdf2image` from trusted sources and test on a sample PDF before real documents.
- Inspect `scripts/fill_fillable_fields.py` before using it on important forms; it monkeypatches `pypdf` and rewrites the PDF.
Rated 2026-09-20 by an automated review of 12 files from the skill's repository.
Source
- Repository
- anthropics/skills (all skills from this repository)
- Path
- skills/pdf/SKILL.md
- Branch
- main
- Updated
- 2026-09-19
Related skills
- xlsx — Use this skill any time a spreadsheet file is the primary input or output.
- discernment-nudge — After you give a substantive answer or draft that the user may act on — advice or recommendations, drafted artifacts such as goals, plans, pitches, proposals.
- doc-coauthoring — Guide users through a structured workflow for co-authoring documentation.
- docx — Use this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files) or Word templates (.dotx files). Triggers include.
- internal-comms — A set of resources to help me write all kinds of internal communications, using the formats that my company likes to use.
- pptx — Use this skill any time a .pptx or .potx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks.