Slack Digest Engine
Scan Slack channels for decisions, action items, and @mentions -- then surface what matters.
Overview
The Slack Digest Engine solves the "200 unread messages" problem. Instead of scrolling through every channel after a meeting, a flight, or a focused work block, you run a single command and get a structured digest that separates signal from noise. Decisions are highlighted, action items are extracted with assignees, your @mentions are listed, and everything else is scored and ranked by importance.
This namespace provides two complementary commands. Catch-up is the lightweight, personal scanner -- it finds only messages that directly need your attention (your @mentions, tasks assigned to you, critical broadcasts) and is designed for speed. Digest is the full team-level scan -- it classifies every message, detects decisions, extracts action items for all team members, and optionally saves the full digest to Notion.
Both commands use a 4-factor scoring algorithm (message type, engagement, recency, channel importance) to assign priority tiers from P1 (critical) to P5 (noise), and a noise filter that typically removes 40-60% of messages before they reach you. The result: you spend 2 minutes reading instead of 20 minutes scrolling.
Required Tools
| Tool | Required | Purpose |
|---|---|---|
| Slack MCP | Yes | Channel scanning, message history, thread resolution |
| Notion CLI | No | Optional storage of digests to the Briefings database |
| Business Context | No | Personalizes output with your company context |
Commands
/founder-os:slack:catch-up
What it does -- Run a fast personal scan across all accessible Slack channels. Shows only messages directly relevant to you: your @mentions, action items assigned to you, and critical @channel/@here broadcasts. Uses approximately 80% fewer API calls than a full digest.
Usage:
/founder-os:slack:catch-up [--since=8h]
Example scenario:
You step out of a 2-hour client meeting and want to know if anything needs your immediate attention. You run
/founder-os:slack:catch-up --since=2hand in about 30 seconds get a clean list: 1 action item assigned to you in #engineering ("review the PR by EOD"), 2 direct @mentions in #product, and zero critical broadcasts. You know exactly what to respond to.
What you get back:
- Your Action Items -- tasks assigned to you with assignee, channel, due date, and thread context, sorted by urgency
- Your @Mentions -- direct mentions with message excerpt, author, timestamp, and thread context, sorted by recency
- Important Broadcasts -- only @channel/@here messages that score P1 (critical), if any exist
- Channel scan summary showing how many channels were checked and how many were skipped
Flags:
| Flag | Default | Description |
|---|---|---|
--since=TIMEFRAME |
8h |
Time window to scan. Accepts Nh (hours), Nd (days), or YYYY-MM-DD |
/founder-os:slack:digest
What it does -- Produce a detailed, structured digest from one or more Slack channels. Classifies every message, detects decisions, extracts action items with assignees, surfaces your @mentions, and ranks key threads by importance. Optionally saves the full digest to Notion.
Usage:
/founder-os:slack:digest [#channel1 #channel2] [--all] [--since=24h] [--include-dms] [--output=notion|chat|both]
Example scenario:
It is Monday morning and you want to know everything important that happened over the weekend across your key channels. You run
/founder-os:slack:digest --all --since=2d --output=bothand get a structured digest covering: 3 decisions made in #product, 5 action items (2 assigned to you), 4 direct @mentions, and the top 10 most important threads across all channels. The digest is also saved to your Notion Briefings database for reference.
What you get back:
- Decisions -- detected decision statements with author, channel, confidence level, and thread link
- Action Items -- extracted tasks split into "Assigned to you" (first) and "Assigned to others," each with assignee, due date, and source channel
- Your @Mentions -- direct mentions with context
- Key Threads -- top 10 high-priority threads not already surfaced in other sections, ranked by signal score
- Channel Summaries -- one-line overview per channel when 3 or more channels are scanned
- Direct Messages -- included only when
--include-dmsis active - Header statistics: messages analyzed, noise filter rate, channels scanned, time window
Flags:
| Flag | Default | Description |
|---|---|---|
#channel1 #channel2 |
-- | Specific channels to scan (strip the # prefix) |
--all |
Off | Scan all bot-accessible channels |
--since=TIMEFRAME |
24h |
Time window: Nh (hours), Nd (days), or YYYY-MM-DD |
--include-dms |
Off | Include direct messages in the scan |
--output=DEST |
both |
Output destination: notion, chat, or both |
--schedule=EXPR |
-- | Schedule recurring digests (e.g., weekdays at 6 PM) |
How Does Slack Digest Prioritize Messages?
Every message is scored using a 4-factor algorithm: message type (decision, action item, question, or broadcast), engagement level (reactions and thread replies), recency, and channel importance. Scores map to priority tiers from P1 (critical) to P5 (noise). A noise filter removes 40-60% of low-signal messages before the digest reaches you, so the output focuses on what actually requires your attention.
Tips & Patterns
- Use catch-up for quick checks, digest for deep reviews. Catch-up answers "do I need to respond to anything?" in 30 seconds. Digest answers "what happened across the team?" in 2-3 minutes.
- Schedule your digest. Set up a recurring end-of-day digest with
--schedule="0 18 * * 1-5"to get a weekday summary delivered automatically. - Start with specific channels. If you only care about #product and #engineering, name them explicitly instead of using
--all. Smaller scope means faster results and less noise. - Adjust the time window. After a long weekend, use
--since=3d. After a quick lunch break,--since=2his plenty. - Save to Notion for your records. Using
--output=bothgives you the digest in chat immediately and stores a copy in your Notion Briefings database for future reference. - Combine with action tracking. When the digest surfaces action items, create follow-up tasks in Notion with
/founder-os:notion:createor track them in your existing project database.
Related Namespaces
- Notion -- Store digests to Notion; create tasks from discovered action items
- Drive -- Slack threads often reference Drive documents; use Drive Brain to find and summarize them
- Client -- Client-related Slack mentions can inform your client dossier
- CRM -- Decisions and commitments from Slack may need to be logged as CRM activities