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
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 microsoft/cat-agent-skills --skill what-to-use-whengit clone --depth 1 https://github.com/microsoft/cat-agent-skillsWrote 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/what-to-use-when)<a href="https://agentmods.dev/skills/microsoft/cat-agent-skills/what-to-use-when"><img src="https://agentmods.dev/badge/skills/microsoft/cat-agent-skills/what-to-use-when/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/what-to-use-when"><img src="https://agentmods.dev/badge/skills/microsoft/cat-agent-skills/what-to-use-when.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00258 | $0.02854 |
| Opus 5 | $0.00129 | $0.01427 |
| Sonnet 5 | $0.00052 | $0.00571 |
| Haiku 4.5 | $0.00026 | $0.00285 |
Grade A, and why
what-to-use-when 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 6d 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 — 198 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Overview
Cowork runs on usage-based consumption credits on top of the Microsoft 365 Copilot license. Most everyday single-output tasks can be done instead by a 1P Copilot capability that is included in the M365 Copilot license at no extra cost. This skill's job is to guide the user to that free capability — classify the request, recommend the right 1P agent, and only proceed in Cowork if the user explicitly insists. The intent is to minimize Cowork spend and train users to reach for M365 Copilot for these tasks.
Precedence — run this BEFORE acting (most important rule)
This gate runs before any thinking, retrieval, TaskCreate, file generation, or
invoking a doc/deck/spreadsheet skill (pptx / docx / xlsx). For an eligible
request it takes priority over action-bias and gather-before-generate: if you
catch yourself about to reason through, search, create a task, or start building a
deck/doc/sheet for a single-output request (e.g. "create a presentation"),
STOP and run Phase 1–3 below (classify → recommend → confirm) first. Only
retrieve data or build anything after the user explicitly chooses an option.
Exclusion check FIRST (before any classification). If the user's message is a direct M365 action — an imperative to send / reply / forward an email, schedule / cancel / decline / reschedule a meeting, post to a Teams channel, or create a calendar event — this gate does NOT apply. Do not classify, do not show the confirm card. Hand straight back to the normal action path and let it execute. These actions are out of scope even though they produce a single output, because the user is taking an action, not asking for content a free 1P agent could generate instead. Only after a message clears this exclusion do you proceed to Phase 1.
Quick Start
User: "Cowork, summarize my Teams messages from the last 5 weeks."
The user sees a crisp, 2-part reply that points them to the free option:
Classification: This is a quick Teams catch-up — a single summary.
Recommendation: Microsoft 365 Copilot Chat can do this at no extra cost —
it's included in your Copilot license. Just ask it in Copilot Chat.
Then an AskUserQuestion gate:
[Use Copilot Chat — included] / [Do it here in Cowork instead]
Run in Cowork only if the user picks Cowork.
What ships with it
2 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.
- 6d ago First seen · 198 lines · 258 tokens per session scan A 2d853d0c2dd2
what-to-use-when is a skill published in the GitHub repository microsoft/cat-agent-skills (66 stars, last pushed today), licensed MIT. It adds 258 tokens to every session and 2,854 once invoked, about $0.0013 per session on Opus 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
orbit-notion
Open Orbit briefing skill — selected by the Orbit pipeline when Notion is the user's only connected connector, or when the user explicitly scopes their daily digest to Notion. Pulls the past 24 hours of document edits, comments, mentions, and database row changes from the user's authenticated Notion connection and…
Cortex
Operate Cortex, the LifeOS memory system — the typed Knowledge Archive (People, Companies, Ideas, Research with typed related: links) plus recall of prior work sessions, ISAs, and conversations. Search, add, harvest, develop, ingest, distill, graph-navigate, recall. USE WHEN cortex, knowledge, knowledge base, search…
pinchtab-mcp
Use this skill when a task requires browser automation through PinchTab's MCP server connected to a remote browser instance. Covers navigation, element interaction, data extraction, form filling, multi-step flows, and session management via MCP tools.
feishu
Work with Feishu or Lark bots, docs, sheets, bitables, approval flows, and OpenAPI/MCP setup without hardcoding credentials.
peekaboo
Capture and automate macOS UI with the Peekaboo CLI.
mochi-remind
Handle due reminders — notify the user with natural language and mark them done.