Mmk notion database ai summary

Skill magic-meal-kits/mmk-skills/skills/mmk-notion-database-ai-summary

Claude Code skills for Magic Meal Kits CLI — Notion, Paymint, Threads, YouTube automation

Install

npx -y skills add magic-meal-kits/mmk-skills --skill mmk-notion-database-ai-summary

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Bulk AI summary generation across all pages in a Notion database. Triggers on "generate AI summaries for database", "bulk AI summary", "batch AI summary", "run AI summary across database".

SKILL.md

3.5 KB, as published. Nobody here has run it

Recipe: Bulk AI Summary for Notion Database

Prerequisite: Read ../mmk-shared/SKILL.md for auth, global flags, and error handling. Uses: mmk-notion-database — Schema, query, and AI summary commands.

A multi-step workflow that discovers AI configuration for a database, collects all page IDs, and triggers AI summary generation in batches.

This recipe has side effects (triggers AI summary writes on Notion pages). It requires manual invocation only.

Workflow

Step 1: Discover database properties

mmk notion database schema --database-id <db-uuid> -o json

Review the schema output to understand the database structure and confirm it has an AI summary property.

Step 2: Detect AI configuration

mmk notion database schema --database-id <db-uuid> --detect-ai --page-id <page-uuid> -o json

Pick any existing page ID from the database. The output provides:

property-id — the AI summary property identifier
workflow-id — the AI workflow identifier

Both values are needed for the ai-summary command. Save them for Step 4.

Step 3: Collect all page IDs

mmk notion database query --database-id <db-uuid> -o json

Collect all page IDs from the query results. If the database has more than 100 pages, use --cursor for pagination:

mmk notion database query --database-id <db-uuid> --cursor "<next-cursor>" -o json

Repeat until all pages are collected.

Step 4: Trigger AI summary in batches

mmk notion database ai-summary --database-id <db-uuid> --page-id <id-1> --page-id <id-2> ... --property-id "<property-id>" --workflow-id "<workflow-id>" -o json

Limits: Max 10 page IDs per execution. Split collected pages into batches of 10 or fewer.

For each batch:

Run the ai-summary command with up to 10 --page-id flags
Report progress (e.g., "Batch 2/5 complete — 20/47 pages processed")
Continue to the next batch

Step 5: Validate summaries

mmk notion database query --database-id <db-uuid> -o json

Query the database again and check that AI summary fields are now populated across the pages.

Parameters

Parameter	Required	Description
`database-id`	Yes	Notion database ID or URL
`page-id`	Yes (Step 2)	Any existing page ID for AI config detection

Example Session

User: /mmk-notion-database-ai-summary

Claude: Which database would you like to generate AI summaries for?

User: Database ID is abc123-def456. Use page xyz789 to detect the AI config.

Claude: [Step 1] Fetching database schema...
       Found 5 properties: Name, Status, Tags, Summary, Created.

       [Step 2] Detecting AI configuration...
       Found property-id: "prop_abc" and workflow-id: "wf_123".

       [Step 3] Collecting all page IDs...
       Found 23 pages in the database.

       [Step 4] Triggering AI summaries in batches of 10...
       Batch 1/3 complete — 10/23 pages processed.
       Batch 2/3 complete — 20/23 pages processed.
       Batch 3/3 complete — 23/23 pages processed.

       [Step 5] Validating...
       All 23 pages now have AI summaries populated.
       Done.