Leon is an open-source personal AI assistant that uses tools, memory, context, and agent-like execution to understand goals and complete tasks. People use it as a practical assistant that can run locally and work across structured workflows, while the catalogue skills support its operation.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add leon-ai/leon --skill document-authoringgit clone --depth 1 https://github.com/leon-ai/leonWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/leon-ai/leon/document-authoring)<a href="https://agentmods.dev/skills/leon-ai/leon/document-authoring"><img src="https://agentmods.dev/badge/skills/leon-ai/leon/document-authoring/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/leon-ai/leon/document-authoring"><img src="https://agentmods.dev/badge/skills/leon-ai/leon/document-authoring.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00024 | $0.01371 |
| Opus 5.5 | $0.00010 | $0.00548 |
| Sonnet 5.5 | $0.00005 | $0.00274 |
| Haiku 4.5 | $0.00002 | $0.00137 |
Grade A, and why
document-authoring scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 27 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Document Authoring
- Establish the requested format, audience, content and acceptance criteria from the brief and supplied references. Reuse source material; for complex work, track criteria and verification evidence in plan details.
- Choose the creation tool based on the requested output:
- Charts: use
document.echarts.renderwith one named entry per requested chart and native ECharts JSON options. Ground the values in supplied/retrieved data, label units, and select only the requested image formats. Each chart becomes a separate SVG/PNG artifact. Use both formats when a vector document figure and raster preview are useful. Inspect a PNG preview and fix clipping or misleading axes before delivering it. Do not use image-generation models to invent or redraw quantitative charts. - Presentations: use
document.slidev.createfor an editable project, authorslides.mdandstyle.csswith the file tool, and place selected chart/image assets inpublic/. Treat the scaffold as editable: owner choices for theme, layout, colors, typography, icons and animation override all defaults. Replace scaffold styles and components when needed; ensure additional themes have compatible installed dependencies. Use local Lucide icons and coherent cover/section/content layouts. Default to fully visible slides and a simpletransition: fadebetween them. Animate mechanisms inside the slide automatically while it is active: moving evidence along a flow, state changes, process cycles or grounded chart evolution. Use CSS orv-motionwith a static reduced-motion fallback. Avoid generic text/panel entrance and exit effects. Add click-triggered reveals (v-clickor click motion variants) and elaborate slide transitions only when the owner explicitly requests them. Ground technical claims in retrieved context/source and label illustrative data. Batch independent grounding reads and author related files together. Use fenced code or HTML-escaped<pre v-pre><code>for literal examples; keep Vue markup balanced. Runinspect, thencheck; during repairs select affectedoptions.slides. Targeted success hascomplete=falseand cannot establish full-deck coverage. Run the default full check on the final source and requirescope=full,complete=trueandok=true. Batch independent checks/previews and requested exports once source is stable; never edit files while those operations are reading them. Reuse current-run evidence and successful results while relevant source/assets are unchanged. Preview early/late motion frames and, only for requested staged reveals, initial/final click states; a passing check alone does not establish visual or animation quality. Finish withpresentto start/reuse a managed native Slidev server and open the deck in the host’s default browser. Return its audience and presenter links, Space/Right/Left navigation andffullscreen control. Verify opening with the browser tool; ifbrowserOpened=false, explain the reported issue and host-local URL scope. Use Satellite for device-side presentation when Leon runs remotely and ensure the project is on that device. Keep the session until the owner finishes, thenstopits sessionId. Export web/PDF/PNG/PPTX downloads only when requested; opening a deck should not require extracting files or running commands. A web ZIP includesserve.mjs: extract and runnode serve.mjs, then open its localhost URL; it can also be hosted as a static site. PDF/PNG and PowerPoint exports capture states rather than native browser motion. Normal PPTX contains slide images; editable PPTX may rasterize complex elements and needs fidelity review. Shared web decks omit notes by default; an optional source ZIP (includeSource=true) always contains them; omit that archive unless requested. - PDF: write a local
.typsource with the file tool, then calldocument.typst.compile. Keep the source and selected assets together in a project directory. Use the smallestrootPathcontaining that project. Built-in/local assets require no generation-provider account or package download. - Editable Word: use
document.document.createwithformat=docx. Typst does not export DOCX. - SVG/PNG: compile Typst with the requested format. Each selected page becomes an image artifact. For a standalone figure, set the page dimensions to fit the figure rather than exporting a full report page.
- Plain text, CSV, Markdown and source files: use existing file writing.
- Hosted generation or Office formats requiring a provider: inspect
document.document.capabilities, then usegeneratewith the configured defaults. ReservegenerateWithModelfor an explicitly owner-requested provider/model.
- Charts: use
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago Changed 13ab7318004a
- 7d ago Changed · +1 lines · -1 tokens per session 16716813c57b
- 8d ago Changed · +7 lines · +2 tokens per session 2938619fbb6f
- 16d ago First seen · 19 lines · 23 tokens per session scan A 50ab6b305b2d
document-authoring is a skill published in the GitHub repository leon-ai/leon (17,579 stars, last pushed 4d ago), licensed MIT. It adds 24 tokens to every session and 1,371 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-30.
Other skills, from other repositories
wowerpoint
Turn one document into a kawaii NotebookLM slide deck. Use for "wowerpoint this", "make a deck about ", "turn this report into slides", or any request to render a single document as narrative slides.
marimo-pair
Work inside the user's live marimo notebook from the code editor: run Python in the same kernel the user does, inspect live notebook state, and commit durable notebook changes through code mode. Use whenever you create, analyze, or improve the user's marimo notebook.
gpu-document-processing
Use when processing large PDFs, document collections, or bulk text extraction tasks that benefit from GPU-accelerated processing. Triggers when the user provides large documents or needs bulk document analysis.
csv-workbench
Analyze CSV files in /mnt/data and return concise numeric summaries.
xberg
Extract text, tables, metadata, and images from 107 document formats (PDF, Office, images, HTML, email, archives, academic) using Xberg. Use when writing code that calls Xberg APIs in Python, Node.js/TypeScript, Rust, or CLI. Covers installation, extraction (sync/async), configuration (OCR, chunking, output format)…
xberg-extracting-tables
Use when extracting tabular data from PDFs, spreadsheets, or images. Covers layout-aware table detection, table model selection, output formats (markdown / JSON cells), and known limits.