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 action-items-todogit 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/action-items-todo)<a href="https://agentmods.dev/skills/microsoft/cat-agent-skills/action-items-todo"><img src="https://agentmods.dev/badge/skills/microsoft/cat-agent-skills/action-items-todo.svg" alt="Measured on agentmods" 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 | $0.00139 | $0.03160 |
| Opus 5 | $0.00069 | $0.01580 |
| Sonnet 5 | $0.00028 | $0.00632 |
| Haiku 4.5 | $0.00014 | $0.00316 |
Grade A, and why
action-items-todo 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 8d 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 — 171 lines — stays where its author put it; the contents beside it link to each section on GitHub.
:: ACTION ITEMS TO MICROSOFT TO DO :: Monitors Teams chats, meeting transcripts, and Outlook mail for action items EXPLICITLY directed at the user, and captures each one as a Microsoft To Do task with a due date, importance, and owner tagging. Microsoft To Do is the only task store this skill writes to.
Nothing about the user (name, email, list name, schedule, customer and workstream names) is hardcoded. Everything lives in the config file written by SETUP.
== FILES == Both live in the user's home directory, so write them with the platform's own separator and never hardcode a Windows path:
- CONFIG: /.scout/action-items-todo/config.json
- STATE (dedupe + last scan): /.scout/action-items-todo/state.json Resolve at runtime (Node os.homedir(), $HOME, or %USERPROFILE% on Windows). Create the directory if missing. Write both files as UTF-8 JSON, atomically (temp file then rename). Always preserve unknown fields.
== CONFIG SCHEMA == { "version": 1, "identity": { "displayName": str, "mail": str, "upn": str, "aadId": str }, "listName": str, // the Microsoft To Do list tasks are written to "frequency": str, // natural-language interval, e.g. "every 30 minutes" "schedule": "24/7" | "weekdays-allhours" | "workhours" | str, "timeZone": str, // IANA zone name, e.g. "Europe/Berlin" "language": "en", // task titles are always written in this language "owners": { "customers": [str], "workstreams": [str] }, // optional, may be empty "excludedChats": [str], // chat/channel names to always ignore "notifyTeams": bool, // send a Teams self-message when items are captured "automationId": str|null, "setupCompleted": ISO8601|null }
== STATE SCHEMA == { "last_scan": ISO8601, "seen": [ { "key": str, "title": str, "created": ISO8601, "source": "teams"|"transcript"|"email", "url": str|null } ] } "key" is a stable hash of source + thread/message id. Keep the last 500 entries; drop older ones.
== SETUP (first run, or on demand) == Run SETUP automatically whenever CONFIG is missing or setupCompleted is null. Never run a scan before SETUP completes.
- Call workiq_get_my_profile and fill identity. If it fails, say so and stop; the skill cannot target asks without knowing who the user is.
- Read the machine time zone as an IANA name with a cross-platform runtime API: Intl.DateTimeFormat().resolvedOptions().timeZone, e.g.
node -p "Intl.DateTimeFormat().resolvedOptions().timeZone". Do NOT use PowerShell Get-TimeZone: on Windows it returns a Windows zone ID such as "W. Europe Standard Time", which is not an IANA name and will not match the format the rest of the skill expects. If the lookup fails, ask the user for their IANA zone rather than guessing. - Read the user's existing lists with workiq_list_task_lists, then ask QUESTION 1 with m_ask_user, free-text mode, inputHint "list name, e.g. Work": "Which Microsoft To Do list should captured action items go into?" Show the existing list names in the assistant message BEFORE the call so the user can pick one or name a new one. After the reply: if the list does not exist, confirm creation, then workiq_create_task_list.
- Ask QUESTION 2 with m_ask_user, multiple choice, recommended index 1: "How often should I scan for new action items?" Answers: "Every 15 minutes" / "Every 30 minutes" / "Every hour" / "Every 2 hours". Store the matching natural-language interval in frequency.
- Ask QUESTION 3 with m_ask_user, multiple choice, recommended index 0:
"When should the scan run?"
Answers:
- "Around the clock, every day" (desc: best when you work across time zones) -> schedule "24/7"
- "Around the clock, weekdays only" -> "weekdays-allhours"
- "Working hours only, weekdays" (desc: 8am-6pm in your local time zone) -> "workhours"
- "Let me describe it" -> free-text follow-up, stored verbatim.
- Ask nothing else. Default the remaining config: language "en", owners empty, excludedChats empty, notifyTeams true.
- Write CONFIG and an empty STATE with last_scan = now minus one frequency interval. Do NOT seed last_scan further back to "catch up". The first run would then backfill items that were already handled or already completed elsewhere, and create tasks for them. A short first window is correct.
- Create the automation with m_create_automation, then store its id in automationId:
- name: "Action Items to To Do" // avoid > and < in the name; they get HTML-escaped in the automation list
- schedule: derived from frequency + schedule (see SCHEDULE MAPPING)
- teamsNotify: "auto"
- prompt: "Load the action-items-todo skill with m_get_skill and run the SCAN procedure exactly as written. Output only the strict result format the skill defines."
- Confirm back in one short block: list name, frequency, schedule, automation id, and how to change any of them ("run /action-items-todo setup again").
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.
- 8d ago First seen · 171 lines · 139 tokens per session scan A 6fc29055170e
action-items-todo is a skill published in the GitHub repository microsoft/cat-agent-skills (64 stars, last pushed yesterday), licensed MIT. It adds 139 tokens to every session and 3,160 once invoked, about $0.0007 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 2026-08-30.
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…
feishu
Work with Feishu or Lark bots, docs, sheets, bitables, approval flows, and OpenAPI/MCP setup without hardcoding credentials.
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.
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.