17 action items from 9 meetings, 6 are yours
6 of 17 action items are yours; two have a date before Monday. 3 items came out of meetings without an owner.
Every Friday afternoon: the action items from this week's Granola notes, grouped by owner, with the meeting each one came from.
Reads
You get
The card on the right is a sample. Yours shows your own numbers and updates in place every run.
What lands on your board
6 of 17 action items are yours; two have a date before Monday. 3 items came out of meetings without an owner.
Granola writes good notes, and each meeting's next steps sit at the bottom of its own note. By Friday that is a dozen notes, and the follow-ups you promised on Tuesday are buried in the third one. Nobody rereads a week of notes to make a list.
The agent does. It reads every note from this week through Granola's MCP server or API, pulls out the action items, merges the ones that came up in two meetings, and posts them grouped by owner. Yours come first, with the meeting each came from, so the weekend starts with a short list instead of a vague worry.
With the Granola MCP server, call list_meetings for this week; with the API, GET https://public-api.granola.ai/v1/notes?created_after=<Monday>&page_size=30, following cursor while hasMore is true.
MCP: get_meetings for the notes. API: GET /v1/notes/{note_id}, which returns the summary, attendees and calendar event. Transcripts are not needed for this skill.
Take items from next steps or action item sections, and clear commitments in the summary ("Dana will send the revised quote"). Record owner (a named person, or unassigned), the action in one line, any due date mentioned, and the meeting title and date.
Merge items that repeat across meetings, keeping the latest wording. Group by owner with your own items first, then people inside your company, then external attendees.
One line per item with owner, action, due date if any, and the meeting. Note how many meetings had no action items, which is useful too.
With BRIEFKIN_TOKEN set, the agent sends the result as one card that updates in place every run. Without it, you get the report in the chat.
Pick your agent, where the data comes from and how to schedule it; the commands below follow.
One command. It puts the SKILL.md in ~/.claude/skills and in ~/.agents/skills, which Codex, Cursor, Gemini CLI, Copilot CLI and OpenCode read too.
shellcurl -fsSL https://briefkin.dev/skills/granola-meeting-action-items/install | shAsk Claude Code:
promptRun the granola-meeting-action-items skillOr paste the prompt instead of installing:
promptUsing Granola's MCP server (list_meetings and get_meetings) or the Granola API (GET /v1/notes with created_after this Monday, then each note), read all my meeting notes from this week. Pull out action items from next steps sections and clear commitments in the summaries, with owner, the action in one line, any due date mentioned, and the meeting title and day. Merge items repeated across meetings. Group by owner with my items first, then teammates, then external people, and list unassigned items separately. Start with counts: meetings, action items, and how many are mine.Fridays at 16:00, from cron on any machine where Claude Code is signed in:
crontab0 16 * * 5 claude -p "Run the granola-meeting-action-items skill"Add it with crontab -e. A laptop asleep at that time runs it when it wakes if you use launchd or Task Scheduler instead; GitHub Actions works for the API route.
Connect Briefkin once with curl -fsSL https://briefkin.dev/install | sh and every run lands on your board as a report card. How setup works
SKILL.md---
name: granola-meeting-action-items
description: "Every Friday afternoon: the action items from this week's Granola notes, grouped by owner, with the meeting each one came from. Use when the user asks for this, or when it runs on a schedule."
---
# Granola meeting action items
Every Friday afternoon: the action items from this week's Granola notes, grouped by owner, with the meeting each one came from.
Granola writes good notes, and each meeting's next steps sit at the bottom of its own note. By Friday that is a dozen notes, and the follow-ups you promised on Tuesday are buried in the third one. Nobody rereads a week of notes to make a list.
The agent does. It reads every note from this week through Granola's MCP server or API, pulls out the action items, merges the ones that came up in two meetings, and posts them grouped by owner. Yours come first, with the meeting each came from, so the weekend starts with a short list instead of a vague worry.
## Two ways to get the data
Use the API or MCP route when you have an MCP server, a CLI or an API token for the app. Otherwise, if you can drive a browser the user is signed in to (Claude in Chrome, a Playwright or Chrome DevTools MCP server), use the browser route. Say which one you used.
## Route A: API or MCP server
### What you need
- Granola's MCP server (`https://mcp.granola.ai/mcp`, sign in with OAuth). The free plan has limited meeting history, which still covers a weekly run; Business adds unlimited history and shared notes.
- Or a Granola API key (Business or Enterprise plan, created in the desktop app under Settings > Connectors > API keys with the Personal notes scope) in `GRANOLA_API_KEY`, sent as a Bearer token.
If something is missing, say what and stop; do not invent numbers.
### Steps
1. **List this week's meetings** (Granola): With the Granola MCP server, call `list_meetings` for this week; with the API, `GET https://public-api.granola.ai/v1/notes?created_after=<Monday>&page_size=30`, following `cursor` while `hasMore` is true.
2. **Read each note** (Granola): MCP: `get_meetings` for the notes. API: `GET /v1/notes/{note_id}`, which returns the summary, attendees and calendar event. Transcripts are not needed for this skill.
3. **Pull out the action items**: Take items from next steps or action item sections, and clear commitments in the summary ("Dana will send the revised quote"). Record owner (a named person, or unassigned), the action in one line, any due date mentioned, and the meeting title and date.
4. **Merge and group**: Merge items that repeat across meetings, keeping the latest wording. Group by owner with your own items first, then people inside your company, then external attendees.
5. **Post the list**: One line per item with owner, action, due date if any, and the meeting. Note how many meetings had no action items, which is useful too.
## Route B: the browser
### What you need
- Signed in to Granola on the web with the same account as the desktop app, in the workspace whose notes you want; your plan's meeting history must reach back to Monday.
If something is missing, say what and stop; do not invent numbers.
### Steps
Start at https://notes.granola.ai/. The user is signed in already; never type passwords or codes, and ask if a sign-in page appears.
1. **Open Granola on the web**: Granola's web app lists your notes with the newest first. Open the **Home** or **My notes** list and read down to Monday of this week; note each meeting's title, date and attendees.
2. **Read each note's summary**: Open every meeting from this week and read the enhanced notes, especially sections titled **Next steps** or **Action items**, plus any clear commitment in the summary ("Dana will send the revised quote"). Transcripts are not needed.
3. **Record owner, action and date**: For each item, write down the owner (a named person, or unassigned when the note says "we will"), the action in one line, any due date mentioned, and the meeting title and day.
4. **Merge and group**: Merge items that repeat across meetings, keeping the latest wording. Group by owner with yours first, then people inside your company, then external attendees, and list unassigned items separately.
5. **Write the list**: Counts first (meetings, items, how many are yours), then one line per item with owner, action, date and meeting, then the meetings that produced no action items.
Note: The web app shows notes your account can open; notes a teammate has not shared are not there. If a meeting has no enhanced notes yet, there are no action items to read. Granola's own **Ask** (chat across notes) can list action items for the week too, but check its answer against the notes before trusting owners.
## The report
- **Counts**: meetings this week, action items, and how many are yours.
- **Your items** first, then each other owner's, one line each with the meeting.
- **Unassigned** items that need an owner.
- Meetings with no action items, in one line.
## Post it to Briefkin
When `BRIEFKIN_TOKEN` is set, or `~/.config/briefkin/env` exists (the installer from https://briefkin.dev/docs writes it), also send the result to the user's Briefkin board as one card. Keep the `item_key` the same on every run so the card updates in place, use a new `event_id` per run, and put today's real numbers in place of the example ones. If the request fails, say so once and carry on.
```bash
[ -n "${BRIEFKIN_TOKEN:-}" ] || . ~/.config/briefkin/env
curl -sS -X POST "${BRIEFKIN_URL:-https://briefkin.dev}/v1/events" \
-H "authorization: Bearer $BRIEFKIN_TOKEN" \
-H "content-type: application/json" \
-d @- <<'JSON'
{
"v": 1,
"event_id": "granola-meeting-action-items/<date of this run, e.g. 2026-10-16>",
"item_key": "granola-meeting-action-items",
"project": "Meetings",
"source": "Granola meeting action items skill",
"kind": "update",
"title": "17 action items from 9 meetings, 6 are yours",
"summary": "**6 of 17** action items are yours; two have a date before Monday. **3 items** came out of meetings without an owner.",
"widget": [
{
"type": "Table",
"props": {
"columns": [
"Owner",
"Action",
"Meeting"
],
"rows": [
[
"You",
"Send Northwind the revised quote (by Mon)",
"Tue, Northwind pricing"
],
[
"You",
"Share the hiring plan with Priya (by Fri)",
"Wed, leadership sync"
],
[
"You",
"Intro Sam to the Globex CTO",
"Thu, Globex check-in"
],
[
"Dana",
"Draft the renewal terms for Halcyon",
"Tue, Northwind pricing"
],
[
"Ravi",
"Estimate the CSV export fix",
"Wed, product review"
],
[
"Unassigned",
"Decide who owns the partner webinar",
"Thu, marketing sync"
]
]
}
},
{
"type": "Badges",
"props": {
"items": [
{
"label": "Yours 6",
"tone": "accent"
},
{
"label": "Team 8"
},
{
"label": "Unassigned 3",
"tone": "warn"
}
]
}
}
]
}
JSON
```
Widgets come from the Briefkin catalog (https://briefkin.dev/v1/catalog). Without a token, just give the user the report.
## Run it on a schedule
Suggested: Fridays at 16:00 (`0 16 * * 5`). From cron, with the agent signed in on that machine:
```bash
0 16 * * 5 claude -p "Run the granola-meeting-action-items skill"
# Codex: codex exec, Gemini CLI: gemini -p, Copilot CLI: copilot -p, OpenCode: opencode run
```
Or use the agent's own scheduler (Claude Code routines, Codex automations, Claude or ChatGPT scheduled tasks): https://briefkin.dev/skills/granola-meeting-action-items#schedule
Source: https://briefkin.dev/skills/granola-meeting-action-items
Granola's MCP server (https://mcp.granola.ai/mcp, sign in with OAuth). The free plan has limited meeting history, which still covers a weekly run; Business adds unlimited history and shared notes.
Or a Granola API key (Business or Enterprise plan, created in the desktop app under Settings > Connectors > API keys with the Personal notes scope) in GRANOLA_API_KEY, sent as a Bearer token.
https://mcp.granola.ai/mcp, signed in through the browser with OAuth. In Claude, connect Granola from Settings > Connectors. In Claude Code, run claude mcp add granola --transport http https://mcp.granola.ai/mcp, then /mcp and Authenticate. Codex, Cursor and other clients can add the same URL if they support Streamable HTTP and OAuth discovery (docs).The server only reads; on Enterprise, admins decide which notes members can expose to AI tools and can turn transcripts off for the whole workspace while leaving summaries available.
The notes your Granola account can see in the active workspace, within your plan's meeting history (limited on the free plan, unlimited on Business and Enterprise). The API needs a Business or Enterprise plan, follows the key's scopes (personal notes, public notes or both), and returns only notes that have a summary and transcript.
From the notes: Granola's summaries usually name who will do what. When a note says "we will" or names nobody, the item is listed as unassigned rather than guessed, so the list never hands you work that was someone else's.
No. The skill runs with your own agent and gives you the report in the chat. With a Briefkin token set, each run also posts it to your board as a report card, so you can glance at it instead of asking again.
Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot CLI, OpenCode, Windsurf and Amp all read SKILL.md files. The install command puts it in ~/.claude/skills and ~/.agents/skills, where they look. ChatGPT, the Claude app and Grok take the prompt instead. Pick your agent under "Install and run it" and the commands change to match.
Yes. Switch the data source to "Your browser" and the steps change to the Granola pages themselves: the agent opens notes.granola.ai in a browser you are signed in to and reads the numbers off the page. Claude Code, Codex, Cursor, Claude app and ChatGPT can drive a browser out of the box; Gemini CLI, Copilot CLI and OpenCode need a Playwright or Chrome DevTools MCP server first. It is slower than the API and needs you signed in, so it suits a report you run by hand or from a scheduler on your own computer.
Skills are free to use with any agent. Briefkin is where their results land: one board for every report, number and question your agents send, with a ping when something needs you.
Free during the open beta. No card.