More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Markdown-to-pdf

    2

    Chinese or English Markdown with math, tables and code, typeset to A4 by one Python file and one static binary. No Office, no browser, no TeX.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    Markdown-to-pdf

    Example session with this skill installed

    Use my Agensi skill "md-to-pdf" to turn this into a print-ready PDF.

    File: report.md — a physics lab write-up with formulas, a 10-row measurement
    table and one python block.

    Title: Newton's Rings
    Subtitle: Physics Lab 4.18

    A4, with page numbers, please. If it needs anything installed, check first and
    tell me before downloading.

    • Read your context and instructions
    • Compiled the markdown-to-pdf
    • Generated the document

    The result (what your agent delivers)

    
    ## Summary
    
    Converted `report.md` into a 2-page A4 PDF — formulas typeset as vector math,
    the measurement table intact, page numbers stamped. One command, no browser,
    no Office, nothing downloaded.
    
    ## What ran
    
    ```bash
    python md2pdf.py report.md out.pdf "Newton's Rings" "Physics Lab 4.18"
    
    # PDF -> /work/out.pdf  (lang: en, bytes: 121451)
    

    What came out

    ElementResult
    Page geometryA4, 595 × 842 pt, margins 2.4/2.2/2.5/2.6 cm
    Typography12 pt, 1.72 line height, justified, Times-class serif, hyphenation on
    Math2 display + 8 inline equations, native Typst math (vector, selectable, not images)
    Table10 rows × 5 columns, bold header, alignment preserved
    Codepython fence, syntax-highlighted, kept whole (not split across a page)
    Page numbers9 pt, centred: "1", "2"
    FontsTimesNewRomanPSMT, Arial-BoldMT headings, NewCMMath for the maths — all embedded

    Verified, not assumed

    pages: 2   geometry: 595.3 x 841.9 pt   bytes: 121451   exit code: 0
    

    All ten table rows read back out of the PDF, all five section headings found,
    the python block's text extractable, and each page's last line is its own page
    number. That is the check I run before handing a PDF over — a PDF that merely
    "was written" is not a finished job.

    Next steps

    • out.pdf is ready to submit; the source report.md is untouched.
    • Same command, Chinese document: the language is detected from the text, so a
      lab report in Chinese gets 宋体, a 2em first-line indent and CJK punctuation
      rules with no extra flags (--lang overrides if the mix confuses it).
    • If typst is missing on your machine, the script exits 3 and names the
      download — it never produces a half-broken PDF, and never fetches anything
      without asking.
    • --doctor lists typst, Latin serif, CJK serif and monospace availability in
      one table before you start.
    
    ## Notes for whoever fills the form
    
    - Both fields render Markdown, so the tables and the bash fence will look like
      the listing page preview. Keep the backticks exactly as they are.
    - If the editor accepts images, upload `docs/img/en-note-page1.png` right after
      the "What came out" table — it is the same document.
    - The result is ~1,900 characters, comfortably over the 700-character minimum,
      and it is real output: re-run `agensi/md2pdf.py` on `examples/en-note.md` and
      you get the same 121,451 bytes.
    

    out.pdf

    PDF · document

    Generated

    Example file from a real run - the skill writes it into your workspace.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Requires: python3 (3.9+, verified on 3.14) and typst (0.13+, one static binary, ~30 MB). Chinese documents also need a CJK font — built into Windows, apt install fonts-noto-cjk on Linux. If typst is missing the script says so and exits 3 instead of producing a broken PDF; --doctor reports what is present in one table.

    What it does

    • Fixed A4 geometry — 2.4/2.2/2.5/2.6 cm margins, 12 pt, 1.72 line height, justified text, 9 pt centred page numbers.
    • Bilingual by detection — Chinese gets a 2em first-line indent and a CJK serif; English gets no indent, hyphenation and a Times-class serif. One command, no flags; --lang overrides.
    • Math — \( … \) inline, \[ … \] block, typeset by Typst's native math engine. Unsupported TeX commands are printed verbatim and listed on stderr, so a wrong symbol is visible rather than silent.
    • Code and tables — fenced blocks are syntax-highlighted and never split across a page; pipe tables keep their alignment.
    • Deterministic — the same .md produces the same PDF on every run.

    How it runs

    1. python md2pdf.py --doctor — checks typst, a Latin serif, a CJK serif and a monospace face.
    2. python md2pdf.py report.md report.pdf "Title" "Subtitle" — detects the language, converts, compiles. A two-page report takes about half a second.
    3. Exit codes mean something: 3 = typst missing, 4 = typst rejected the document, 5 = nothing was written. --keep-typ keeps the generated Typst source for debugging.

    Why not just ask an agent to do it?

    You get a different pipeline every time — Pandoc here, WeasyPrint there, a --pdf-engine flag that does not exist, formulas that silently become $x$, CJK that renders as boxes. This is one deterministic mapping: Markdown → Typst → PDF, with typographic values that were measured rather than guessed, and two rendered examples in the box to prove it.

    Good for

    • Reports and handouts produced on a headless server or in CI.
    • Technical notes with formulas and code that must look the same in 2027.
    • Bilingual documents where Chinese paragraphs need real indentation.

    Limits (please read)

    Nested lists, footnotes, HTML blocks and images are not supported. Images are not dropped silently — they become a visible [image: alt] marker plus a warning. Math covers a mapped subset; anything else is passed through verbatim.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Writing & Content.

    No support for nested lists, footnotes, HTML, or images. Requires the typst binary. Math requires specific delimiters. CJK rendering depends on system font availability.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 11 days ago

    Software
    Pythontypst

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Environment Variables
    Listed11 days ago

    What's inside

    Frequently Asked Questions

    Traditional PDF conversion often requires gigabytes of TeX libraries or heavy headless browsers; this skill uses a lightweight Typst backend to deliver print-quality A4 documents with a single Python script.

    Yes, it features a specialized detection engine that automatically applies 2em paragraph indents and optimized line heights for CJK text while maintaining professional typography for English.

    The skill requires Python 3.9+ and the Typst CLI binary installed on your system; it includes a built-in 'doctor' mode to quickly verify your environment and font availability.

    You receive the standalone Python bridge script, documentation for CLI integration, and a diagnostic utility to ensure your system is ready for high-performance typesetting.

    The skill handles complex LaTeX formulas for both inline and block equations, though it focuses on standard commands and does not currently support custom TeX macros or external images.

    This skill is designed as a deterministic tool that AI agents can call via CLI to prevent formatting hallucinations, making it compatible with any agent framework that supports local script execution.

    New in Writing & Content