microsoft/cat-agent-skills is a static website that catalogs reusable instruction sets and related packages for AI agents. People use it to search, filter, rate, and download skills for Cowork, Copilot Studio, and Scout, along with Copilot plugins and Scout automations. The catalogue entries are the skills, instructions, plugins, and settings displayed by the site.
Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/microsoft/cat-agent-skillsnpx agentmods add skills/microsoft/cat-agent-skills/scrollytelling-dataWrote 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/microsoft/cat-agent-skills/scrollytelling-data)<a href="https://agentmods.dev/skills/microsoft/cat-agent-skills/scrollytelling-data"><img src="https://agentmods.dev/badge/skills/microsoft/cat-agent-skills/scrollytelling-data/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/microsoft/cat-agent-skills/scrollytelling-data"><img src="https://agentmods.dev/badge/skills/microsoft/cat-agent-skills/scrollytelling-data.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Prompt Injection · line 186 Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.Fix: Audit all comments and invisible characters. Remove any instructions that direct the agent to perform unauthorized actions. Use plain, reviewable content.
- high Prompt Injection · line 371 Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.Fix: Audit all comments and invisible characters. Remove any instructions that direct the agent to perform unauthorized actions. Use plain, reviewable content.
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 | $0.00107 | $0.10052 |
| Opus 5.5 | $0.00043 | $0.04021 |
| Sonnet 5.5 | $0.00021 | $0.02010 |
| Haiku 4.5 | $0.00011 | $0.01005 |
Grade A, and why
scrollytelling-data 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 1mo 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 — 608 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill: Data Scrollytelling
Turn an uploaded dataset (.xlsx, .csv, .json, or PDF tables) into a single self-contained scroll-driven HTML data story: animated Plotly charts, counting stat cards, and narrative chapters. Plotly.js loads from a CDN at view time.
How to Talk to the User While Working
The internal workflow uses technical steps (SQL, JSON, HTML, Plotly), but never expose that vocabulary in chat messages. Narrate like a friendly analyst explaining to a non-technical stakeholder — plain language, short sentences, a little emoji. Never say "SQL," "JSON," "HTML," "Plotly," "dataframe," or "aggregation" in a user-facing message.
Send a checkpoint per meaningful sub-step, not one per phase. Phase 1 (inspect) and Phase 3 (aggregations) can each run 30–60+ seconds silently, which feels stalled — break them into several short updates as work happens. Don't number them (no Step X/Y); lead with an emoji and a short, plain-language title.
Example checkpoints (an illustrative full run — drop any line that doesn't apply, e.g. no time or geo column, and split a slow phase into more):
- 👀 "Taking a first look at your data..."
- 📐 "Found
<N>rows and<N>columns — checking what each one means..." - 🏷️ "Spotted your key categories:
<examples>..." - 🔢 "Counting up the totals..."
- 🏆 "Ranking your top performers — your #1 is
<top entity>..." - 🥧 "Breaking things down by category..."
- 📈 "Mapping out the trend over time..." (skip if no time column)
- 🗺️ "Plotting things out by location..." (skip if no geo column)
- ✍️ "Writing the story chapters..."
- ✅ "Your story is ready! Take a look 👇"
Give ranking and chapter-writing extra beats — that's where the user most wants to feel progress. Fold a real finding into the update (the actual #1, the size of the gap, the standout insight) with real names and numbers ("🥇 Sales leads with 3,120 hours saved..."), not placeholders. Keep every line short — a few words plus one emoji — and save the substance (findings, chapter takeaways) for the delivered story itself.
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago First seen · 608 lines · 107 tokens per session scan A 1d6d9842b4f0
scrollytelling-data is a skill published in the GitHub repository microsoft/cat-agent-skills (93 stars, last pushed today), licensed MIT. It adds 107 tokens to every session and 10,052 once invoked, about $0.0004 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 2026-09-03.
Other skills, from other repositories
doc-reader
Read any common document/data file — PDF, Word (.docx), Excel (.xlsx/.xls), PowerPoint (.pptx), images (OCR), CSV/TSV, plain text, JSON/YAML/TOML, HTML/XML, and most source-code files. Use the readdocument tool.
surfsense-documents
Make or change a Word document (.docx), a PDF, a PowerPoint deck (.pptx) or an Excel workbook (.xlsx) by writing a Python script that surfsenserenderdocument runs, optionally starting from a source's own Word or PowerPoint file. Load it whenever the user asks for one of these, or for a change to one you made.
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.
skill-doc-delivery
Convert markdown to DOCX, PPTX, XLSX, PDF office documents — use when you need exportable deliverables.
report-generator
A report-generation tool for producing SEO and GEO analysis reports in formats such as Markdown, HTML, PDF, JSON, and Excel. It also supports charts, templates, data processing, and interactive report elements.
kordoc
A command-line and server tool for reading and creating Korean official documents, including HWP, HWPX, PDF, DOCX, and spreadsheet files. It can convert documents to Markdown and create or update HWPX files.