How do I install the PDF To Markdown skill?

Run npx skillmds@latest add to-aoki/pdf-to-markdown 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 To Markdown safe to use?

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

Which AI agents work with PDF To Markdown?

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 To Markdown free to use?

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

Who published PDF To Markdown?

to-aoki (@to-aoki) published this skill. Their other Agent Skills are listed on their SkillMD profile.

PDF To Markdown by to-aoki — Agent Skill

Name: PDF To Markdown
Rating: 4.5 (4 reviews)
Author: to-aoki

PDF To Markdown

Use this skill when the user asks to convert a PDF file into Markdown. The file path may be explicit, partial, or omitted if the target PDF can be inferred from the workspace. Use it for text-based .pdf files and create or return Markdown output by running the bundled pdfplumber-based script.

to-aoki Updated Sep 22, 2026

File contents

PDF to Markdown

Use this skill only for PDF to Markdown conversion. For every conversion request, you must use the bundled script scripts/pdf_to_markdown.py. Do not convert PDF files manually in your own reasoning or by writing the Markdown yourself.

Supported files

.pdf

Use this skill for text PDFs. If the PDF appears to be image-only or scanned, say that plain text extraction may be limited because this skill does not perform OCR.

Do This

Find the target PDF file.
Run uv run scripts/pdf_to_markdown.py once. This step is mandatory.
If the user asked for a file, use --output.
If the user did not ask for a file, return the Markdown result.
After the conversion result is available, stop.

Workspace discovery and output paths must stay inside the workspace scope provided by the host runner.

Do Not Do This

Do not think in loops.
Do not repeat the same judgment.
Do not output self-talk.
Do not run the script more than once unless the first run clearly failed.
Do not do extra formatting passes.
Do not answer from memory instead of running the script.
Do not inspect PDF bytes and then compose the Markdown by hand.
Do not skip the script even if the expected output seems obvious.

Commands

Basic:

uv run scripts/pdf_to_markdown.py --input <path>

Write file:

uv run scripts/pdf_to_markdown.py --input <path> --output <path-to-markdown>

If the file path is missing, use workspace discovery:

uv run scripts/pdf_to_markdown.py --kind pdf

If the file path is vague, use a query:

uv run scripts/pdf_to_markdown.py --kind pdf --query "quarterly report"

Optional:

--title <title>
--template <path>
--kind <pdf>
--query <text>

The host runner prepends --workspace <path> automatically.

Templates

Default template:

assets/pdf-template.md

Use a custom template only if the user asks for a specific layout.

Output Rules

Preserve the extracted structure page by page.
Keep page headings, paragraphs, and tables when possible.
If a page has no extractable text, say so plainly.
If the whole PDF has little or no extractable text, say so plainly.

Final Response

If you created a file, say which file you created. If you did not create a file, return the converted Markdown. Do not add analysis before or after the result. Before the final response, confirm that the script was run and use its result.

---name: pdf-to-markdowndescription: Use this skill when the user asks to convert a PDF file into Markdown. The file path may be explicit, partial, or omitted if the target PDF can be inferred from the workspace. Use it for text-based .pdf files and create or return Markdown output by running the bundled pdfplumber-based script.---# PDF to MarkdownUse this skill only for PDF to Markdown conversion.For every conversion request, you must use the bundled script `scripts/pdf_to_markdown.py`.Do not convert PDF files manually in your own reasoning or by writing the Markdown yourself.## Supported files- `.pdf`Use this skill for text PDFs.If the PDF appears to be image-only or scanned, say that plain text extraction may be limited because this skill does not perform OCR.## Do This1. Find the target PDF file.2. Run `uv run scripts/pdf_to_markdown.py` once. This step is mandatory.3. If the user asked for a file, use `--output`.4. If the user did not ask for a file, return the Markdown result.5. After the conversion result is available, stop.Workspace discovery and output paths must stay inside the workspace scope provided by the host runner.## Do Not Do This- Do not think in loops.- Do not repeat the same judgment.- Do not output self-talk.- Do not run the script more than once unless the first run clearly failed.- Do not do extra formatting passes.- Do not answer from memory instead of running the script.- Do not inspect PDF bytes and then compose the Markdown by hand.- Do not skip the script even if the expected output seems obvious.## CommandsBasic:`uv run scripts/pdf_to_markdown.py --input <path>`Write file:`uv run scripts/pdf_to_markdown.py --input <path> --output <path-to-markdown>`If the file path is missing, use workspace discovery:`uv run scripts/pdf_to_markdown.py --kind pdf`If the file path is vague, use a query:`uv run scripts/pdf_to_markdown.py --kind pdf --query "quarterly report"`Optional:- `--title <title>`- `--template <path>`- `--kind <pdf>`- `--query <text>`The host runner prepends `--workspace <path>` automatically.## TemplatesDefault template:- `assets/pdf-template.md`Use a custom template only if the user asks for a specific layout.## Output Rules- Preserve the extracted structure page by page.- Keep page headings, paragraphs, and tables when possible.- If a page has no extractable text, say so plainly.- If the whole PDF has little or no extractable text, say so plainly.## Final ResponseIf you created a file, say which file you created.If you did not create a file, return the converted Markdown.Do not add analysis before or after the result.Before the final response, confirm that the script was run and use its result.

Frequently asked questions