How do I install the Markwatch skill?

Run npx skillmds@latest add zurmely/markwatch 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 Markwatch 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 Markwatch?

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 Markwatch free to use?

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

Who published Markwatch?

Zurmely (@zurmely) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Markwatch by zurmely — Agent Skill

Name: Markwatch
Author: Zurmely

Markwatch

Converts documents (PDF, DOCX, PPTX, XLSX, XLS, CSV, HTML, images, and anything else MarkItDown supports) into a Markdown file sitting right next to the original, without touching or deleting the source. Use this whenever the user wants to read, search, summarize, or otherwise work with the text content of a non-Markdown file, wants to convert a document to Markdown, or hands you a file path/attachment whose extension isn't .md and asks you to look at it. Handles installing the markitdown CLI on first use automatically — don't try to install it or shell out to markitdown yourself.

Zurmely Updated Sep 22, 2026

File contents

MarkWatch conversion skill

Companion to the MarkWatch VS Code extension (see the repo's AGENTS.md): same non-destructive conversion, same output naming, but for agent/CLI contexts instead of the editor.

When to use this

Any time you need the text/content of a file that isn't already Markdown or plain text — a .docx, .pdf, .pptx, .xlsx/.xls, .csv, .html, an image, etc. — convert it first, then read the resulting .md file. Don't try to parse binary formats yourself or write your own extraction code.

How to use it

Run the bundled script with the file path(s) as arguments. It does everything in one shot — checks for markitdown, installs it if missing, skips files that are already converted and up to date, runs the conversion, and prints one status line per file. Don't reason through the install/convert steps yourself or write shell commands to replicate this — that's what the script is for.

python3 <skill_dir>/scripts/convert.py path/to/report.docx [more files...]

Output is one line per file, prefixed OK, SKIP, or FAIL:

OK report.docx -> report.docx.md
SKIP notes.docx.md: up to date
FAIL broken.pdf: <markitdown's error>

The converted file is always <original name>.<original extension>.md next to the source (e.g. report.docx → report.docx.md), matching the extension's naming exactly. Exit code is 0 if everything converted or was skipped as up-to-date, 1 if any file failed.

After it prints OK or a pre-existing SKIP, read the .md file (it's the same path as the source with .md appended) to get the content.

Notes

First run on a machine without markitdown installed will take longer (it installs via uv tool install, falling back to pipx, then pip3 --user) — this only happens once.
The source file is never modified or deleted; only the .md sibling is written.
Safe to call on a batch of files in one invocation — pass them all as separate arguments rather than looping and invoking the script per file.

---name: markwatchdescription: Converts documents (PDF, DOCX, PPTX, XLSX, XLS, CSV, HTML, images, and anything else MarkItDown supports) into a Markdown file sitting right next to the original, without touching or deleting the source. Use this whenever the user wants to read, search, summarize, or otherwise work with the text content of a non-Markdown file, wants to convert a document to Markdown, or hands you a file path/attachment whose extension isn't .md and asks you to look at it. Handles installing the markitdown CLI on first use automatically — don't try to install it or shell out to markitdown yourself.---# MarkWatch conversion skillCompanion to the MarkWatch VS Code extension (see the repo's `AGENTS.md`): samenon-destructive conversion, same output naming, but for agent/CLI contextsinstead of the editor.## When to use thisAny time you need the text/content of a file that isn't already Markdown orplain text — a `.docx`, `.pdf`, `.pptx`, `.xlsx`/`.xls`, `.csv`, `.html`, animage, etc. — convert it first, then read the resulting `.md` file. Don't tryto parse binary formats yourself or write your own extraction code.## How to use itRun the bundled script with the file path(s) as arguments. It does everythingin one shot — checks for `markitdown`, installs it if missing, skips filesthat are already converted and up to date, runs the conversion, and prints onestatus line per file. Don't reason through the install/convert steps yourselfor write shell commands to replicate this — that's what the script is for.```bashpython3 <skill_dir>/scripts/convert.py path/to/report.docx [more files...]```Output is one line per file, prefixed `OK`, `SKIP`, or `FAIL`:```OK report.docx -> report.docx.mdSKIP notes.docx.md: up to dateFAIL broken.pdf: <markitdown's error>```The converted file is always `<original name>.<original extension>.md` next tothe source (e.g. `report.docx` → `report.docx.md`), matching the extension'snaming exactly. Exit code is `0` if everything converted or was skipped asup-to-date, `1` if any file failed.After it prints `OK` or a pre-existing `SKIP`, read the `.md` file (it's thesame path as the source with `.md` appended) to get the content.## Notes- First run on a machine without `markitdown` installed will take longer (it installs via `uv tool install`, falling back to `pipx`, then `pip3 --user`) — this only happens once.- The source file is never modified or deleted; only the `.md` sibling is written.- Safe to call on a batch of files in one invocation — pass them all as separate arguments rather than looping and invoking the script per file.

Frequently asked questions