memory-slack-ingest

name	memory-slack-ingest
description	Ingest Slack DMs and important channel messages into the memory knowledge base
user_invocable	true

Slack DM Ingest for Memory Knowledge Base

Pull recent Slack direct messages and important threads, extract decisions and context, append to daily log.

Process

1. Load State

Read state from /Users/jesseanglen/Documents/RandomStuff/memory-kb/state/slack-state.json. If it doesn't exist, this is the first run — set last_checked to 3 days ago (Unix timestamp).

Jesse's Slack user ID: U01TKLJK04B

2. Search Recent DMs

Use slack_search_public_and_private to find recent DMs:

query: "to:me"
channel_types: "im,mpim"
sort: "timestamp"
sort_dir: "desc"
limit: 20

If last_checked exists, use after parameter with the Unix timestamp.

Also search for messages from Jesse in DMs:

query: "from:<@U01TKLJK04B>"
channel_types: "im,mpim"

3. Read Substantive DM Threads

For each DM found, check if it's substantive (not just "thanks" / "ok" / emoji reactions).

For substantive messages, use slack_read_thread with the channel_id and message_ts to get the full conversation context.

What's substantive:

• Questions about deals, clients, or strategy
• Requests for action or decisions
• Updates on prospect status, client work, or internal ops
• Forwarded information or meeting context
• Technical questions or implementation discussions

Skip:

• Single-word replies ("thanks", "ok", "👍")
• Automated bot messages
• Calendar/scheduling confirmations
• Idle chatter

4. Cross-Reference with Calendar Contacts

Check if the DM sender appears in /Users/jesseanglen/Documents/RandomStuff/memory-kb/state/calendar-contacts.json. If they do, note the related_event for context linking.

5. Append to Daily Log

For each substantive DM thread, append to /Users/jesseanglen/Documents/RandomStuff/memory-kb/daily/YYYY-MM-DD.md:

## Slack DM from [Name] @ HH:MM PT
**source:** slack
**from:** [Real Name] (Slack: @username)
**channel_type:** DM | Group DM
**related_event:** [if sender is in calendar contacts, reference their last meeting]

### Key Content
- [Summary of the conversation thread]
- [Decisions made or information shared]
- [Action items or requests]

### Action Items
- [ ] [Any follow-ups identified]

Create daily log with header if it doesn't exist.

6. Update State

Write to /Users/jesseanglen/Documents/RandomStuff/memory-kb/state/slack-state.json:

{
  "last_checked": "Unix timestamp",
  "total_runs": N,
  "last_run": {
    "timestamp": "ISO",
    "dms_found": N,
    "dms_ingested": N,
    "threads_read": N,
    "status": "success"
  },
  "processed_message_ts": ["ts1", "ts2"]
}

Keep last 200 processed message timestamps for dedup.

7. Report

"Slack ingest complete. Found X DMs, ingested Y substantive threads."

MCP Tools Reference

• slack_search_public_and_private — search DMs and channels
• slack_read_channel — read recent messages from a DM (use user_id as channel_id)
• slack_read_thread — read full thread replies
• slack_search_users — look up user details by name
• slack_read_user_profile — get user profile info

Notes

• Uses Slack MCP tools — no browser needed
• Focus on DMs (im) and group DMs (mpim) — not public channels
• Cross-reference with calendar contacts for deal context
• Jesse's user ID: U01TKLJK04B
• Skip bot messages, scheduling confirmations, single-word replies
• For group DMs, capture all participants for context