6 mentions today, 2 still need you
Maya needs sign-off on the pricing page copy by 15:00 tomorrow, and Ravi is waiting on your call about dropping the old export format.
Weekdays at 17:00: every Slack message that mentioned you today, with a link, and whether you replied, so nothing waits overnight.
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
Maya needs sign-off on the pricing page copy by 15:00 tomorrow, and Ravi is waiting on your call about dropping the old export format.
Slack's Activity tab lists mentions, but it mixes the ones you handled with the ones you only glanced at between meetings, and it does not say which threads still wait on you. The cost is the question someone asked at 11:00 that you meant to answer and did not.
At the end of each workday, the agent searches for messages that mentioned you, checks each thread for a reply or reaction from you, and posts the list with the unanswered asks on top. Two minutes before you log off is enough to clear it.
auth.test returns your user_id. If your team is mentioned as a group, list user group IDs with usergroups.list and keep yours.
search.messages with a user token, query: "<@U0123ABCD> after:<yesterday>", sort: "timestamp" and count: 100; repeat for each user group as <!subteam^S0456>. Each match has channel, ts, user, text and a permalink.
For each match, conversations.replies with the channel and the thread's ts. Count it as handled when a later message in the thread is from you, or when you reacted to the message.
Needs you: a question, request or decision addressed to you with no reply. FYI: announcements and thanks. Handled: replied or reacted. Leave out bot messages and your own.
Unanswered asks first with channel, sender, the ask in one line and a link, then FYI, then a count of handled ones.
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/slack-mentions-digest/install | shAsk Claude Code:
promptRun the slack-mentions-digest skillOr paste the prompt instead of installing:
promptUsing the Slack API with my user token (or the Slack connector), find messages from today that mention me: search.messages with my user ID in the form <@U...> and after:<yesterday>, sorted by timestamp, plus mentions of my user groups. For each, check the thread with conversations.replies and count it as handled if I replied later in the thread or reacted to the message. Skip bot messages. List the unanswered questions, requests and decisions first with channel, sender, the ask in one line and a link, then FYI mentions, then a count of handled ones.Weekdays at 17:00, from cron on any machine where Claude Code is signed in:
crontab0 17 * * 1-5 claude -p "Run the slack-mentions-digest 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: slack-mentions-digest
description: "Weekdays at 17:00: every Slack message that mentioned you today, with a link, and whether you replied, so nothing waits overnight. Use when the user asks for this, or when it runs on a schedule."
---
# Slack mentions digest
Weekdays at 17:00: every Slack message that mentioned you today, with a link, and whether you replied, so nothing waits overnight.
Slack's Activity tab lists mentions, but it mixes the ones you handled with the ones you only glanced at between meetings, and it does not say which threads still wait on you. The cost is the question someone asked at 11:00 that you meant to answer and did not.
At the end of each workday, the agent searches for messages that mentioned you, checks each thread for a reply or reaction from you, and posts the list with the unanswered asks on top. Two minutes before you log off is enough to clear it.
## 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
- Slack's MCP server through the Slack connector in Claude, signed in as you; or a Slack app with a user token (`xoxp-`) holding `search:read`, `channels:history`, `groups:history`, `im:history`, `mpim:history` and `usergroups:read`.
If something is missing, say what and stop; do not invent numbers.
### Steps
1. **Find your IDs** (Slack): `auth.test` returns your `user_id`. If your team is mentioned as a group, list user group IDs with `usergroups.list` and keep yours.
2. **Search today's mentions** (Slack): `search.messages` with a user token, `query: "<@U0123ABCD> after:<yesterday>"`, `sort: "timestamp"` and `count: 100`; repeat for each user group as `<!subteam^S0456>`. Each match has `channel`, `ts`, `user`, `text` and a `permalink`.
3. **Check whether you replied** (Slack): For each match, `conversations.replies` with the channel and the thread's `ts`. Count it as handled when a later message in the thread is from you, or when you reacted to the message.
4. **Sort what is left**: Needs you: a question, request or decision addressed to you with no reply. FYI: announcements and thanks. Handled: replied or reacted. Leave out bot messages and your own.
5. **Post the digest**: Unanswered asks first with channel, sender, the ask in one line and a link, then FYI, then a count of handled ones.
## Route B: the browser
### What you need
- Signed in to the Slack workspace at app.slack.com as yourself; any member role works, since the agent sees only channels and DMs you are in.
- If you belong to several workspaces, pick the one to check first; Activity is per workspace.
If something is missing, say what and stop; do not invent numbers.
### Steps
Start at https://app.slack.com/. The user is signed in already; never type passwords or codes, and ask if a sign-in page appears.
1. **Open Activity, filtered to mentions**: Pick the workspace, then click **Activity** in the sidebar. If only **All** and **DMs** show, click **Filters** (or the plus icon) and add **Mentions**. Leave **Unread messages** off: a mention you opened and did not answer still counts.
2. **Search today as a cross-check**: In the search box run `@<your display name> after:<yesterday>`; Slack also offers **Mentions** as a search filter. Activity misses nothing recent, but search also shows mentions of user groups you belong to, with `during:today` as an alternative to `after:`.
3. **Open each mention and check the thread**: Click the mention to jump to the message. It is handled when a later message in the thread or channel is from you, or when your reaction sits under it. Skip bot posts and your own messages.
4. **Copy links and sort**: For each unanswered question, request or decision, use **More actions > Copy link** on the message. Group the rest as FYI (announcements, thanks) or handled.
5. **Write the digest**: Needs you first with channel, sender, the ask in one line and the link, then FYI, then a count of handled mentions.
Note: Activity lists mentions per workspace and keeps a limited history on the free plan, which is fine for one day. It does not say whether you replied, so the agent opens each thread; with more than about 30 mentions a day this gets slow, and the API route is better.
## The report
- **Count** of mentions today and how many still need you.
- **Needs you**: channel, sender, the ask in one line, and a link.
- **FYI**: mentions that need no answer.
- **Handled**: a count of mentions you replied to or reacted to.
## 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": "slack-mentions-digest/<date of this run, e.g. 2026-10-16>",
"item_key": "slack-mentions-digest",
"project": "Slack",
"source": "Slack mentions digest skill",
"kind": "update",
"title": "6 mentions today, 2 still need you",
"summary": "Maya needs sign-off on the pricing page copy **by 15:00 tomorrow**, and Ravi is waiting on your call about dropping the old export format.",
"widget": [
{
"type": "Table",
"props": {
"columns": [
"Channel",
"From",
"Message",
"Status"
],
"rows": [
[
"#launch-q4",
"Maya",
"Can you sign off the pricing page copy by 15:00 tomorrow?",
"Needs you"
],
[
"#eng",
"Ravi",
"Need your call on dropping the v1 CSV export",
"Needs you"
],
[
"#sales",
"Dana",
"FYI: Northwind renewal moved to January",
"FYI"
],
[
"#design",
"Ana",
"Updated onboarding mocks, comments welcome",
"Replied"
],
[
"#general",
"Sam",
"Thanks for the all-hands demo",
"Reacted"
],
[
"Group DM",
"Tom",
"Intro to the Globex team, over to you",
"Replied"
]
]
}
}
]
}
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: Weekdays at 17:00 (`0 17 * * 1-5`). From cron, with the agent signed in on that machine:
```bash
0 17 * * 1-5 claude -p "Run the slack-mentions-digest 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/slack-mentions-digest#schedule
Source: https://briefkin.dev/skills/slack-mentions-digest
Slack's MCP server through the Slack connector in Claude, signed in as you; or a Slack app with a user token (xoxp-) holding search:read, channels:history, groups:history, im:history, mpim:history and usergroups:read.
https://mcp.slack.com/mcp. It only works through Slack apps that Slack has registered, so use the Slack connector in Claude or another approved client, and your workspace admin may need to approve it (docs). It searches messages and files, reads channels and threads, and can send messages.SLACK_BOT_TOKEN (xoxb-) or SLACK_USER_TOKEN (xoxp-). Read scopes: channels:history, channels:read, groups:history for private channels, users:read, and search:read (user tokens only) for search. A bot reads only the channels it was invited to.Leave out chat:write unless the skill should post back into Slack; with history and read scopes alone, the app can't post anything.
It sees what you can see. Search with your user token covers private channels and group DMs you are in, and the history scopes let it check replies there. Direct messages to you are not mentions, so they are not in this list unless someone also @-mentions you.
Slack's search.messages only works with a user token, and a bot sees only channels it was invited to. Slack now points new apps at its assistant.search.context method instead, and the Slack connector in Claude uses Slack's own MCP server; either route signs in as you and needs no bot.
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 Slack pages themselves: the agent opens app.slack.com 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.