Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Compiles multi-source meeting briefing documents by pulling event details from Google Calendar, enriching attendee profiles via Slack and Clay, gathering context from Notion, Slack, and Gmail, and assembling a structured markdown prep document. Triggered when the user asks to prepare for a meeting, brief on attendees, compile meeting context, or get ready for a call.
.claude/skills/0xjitsu-meeting-prep/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 93% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 56% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 10% | 0% |
Multi-source meeting briefing compiler.
Activate this skill when the user:
Use Calendar MCP to fetch the target meeting:
Tool: gcal_get_event (if event ID is known)
Tool: gcal_list_events (if searching by time or title)Extract from the event:
If the user says "my next meeting," use gcal_list_events with timeMin set to now and maxResults: 1.
For each attendee email address, gather profile data in parallel:
Slack profile:
Tool: slack_search_users (search by email)
Tool: slack_read_user_profile (get title, team, timezone, status)Company enrichment:
Tool: find-and-enrich-contacts-at-company (Clay MCP)Extract: company name, size, funding stage, tech stack, recent news.
Recent activity:
Tool: slack_search_public (query: "from:@username" with recent date filter)Surface: what they have been talking about recently, active projects, shared channels.
Fallback: If Clay or Slack enrichment fails for an attendee (external contact), note them as "External — limited profile available" and move on. Do not block the entire workflow.
Run these searches in parallel using keywords from the meeting title and description:
Notion:
Tool: notion-search (query: meeting topic keywords)Pull: related docs, project pages, decision logs, specs.
Slack:
Tool: slack_search_public (query: meeting topic keywords)Pull: recent discussions, decisions, blockers mentioned in channels.
Gmail:
Tool: gmail_search_messages (query: attendee emails + topic keywords)Pull: recent email threads, attachments mentioned, action items from previous correspondence.
Google Drive (if relevant):
Tool: google_drive_search (query: meeting topic keywords)Pull: shared documents, presentations, spreadsheets linked to the topic.
Assemble all gathered data into a structured markdown document:
markdown# Meeting Prep: [Title] **Date:** [Day, Month DD, YYYY] at [HH:MM] [timezone] **Duration:** [estimated or scheduled] **Location:** [Zoom link / Meet link / Physical address] **Organizer:** [Name] --- ## Attendees | Name | Title | Company | Timezone | Notes | |------|-------|---------|----------|-------| | Jane Smith | VP Engineering | Acme Corp (Series B, 120 employees) | PST | Active in #product-launches, discussed Q2 roadmap last week | | John Doe | Product Manager | Internal | EST | Owns the dashboard project, sent budget email on Mar 28 | | External Guest | — | Unknown | — | External — limited profile available | ## Context ### From Notion - **[Project Alpha Spec](notion-link)** — Last updated 3 days ago, covers the feature scope being discussed - **[Q2 OKRs](notion-link)** — Relevant KPIs that may come up ### From Slack - **#product** (2 days ago): Discussion about timeline slip on the dashboard migration - **#engineering** (yesterday): John raised a concern about API rate limits ### From Email - **Thread: "Re: Dashboard Launch Plan"** (Mar 28): Jane requested updated wireframes, John replied with ETA of Apr 2 - **Thread: "Budget Approval"** (Mar 25): Budget was approved with a note to reduce cloud spend by 15% ### From Drive - **[Q2 Roadmap Deck](drive-link)** — Shared by Jane, last edited yesterday --- ## Suggested Talking Points Based on gathered context: 1. **Dashboard migration timeline** — Slack discussion suggests a slip; clarify new ETA 2. **API rate limits** — John flagged this in engineering; discuss mitigation strategy 3. **Budget constraints** — 15% cloud cost reduction target affects infrastructure decisions 4. **Updated wireframes** — Jane is waiting on these; confirm delivery status ## Open Questions - [ ] What is the revised launch date for the dashboard? - [ ] Has the API rate limit issue been scoped? - [ ] Are the wireframes ready for review? - [ ] Who owns the cloud cost reduction initiative? --- *Compiled at [timestamp] by Meeting Prep Skill*
Write the compiled document to:
~/Reports/meeting-prep/YYYY-MM-DD-<meeting-title-slug>.mdRules for the file path:
~/Reports/meeting-prep/2026-04-05-q2-product-roadmap-sync.mdCreate the ~/Reports/meeting-prep/ directory if it does not exist.
For recurring prep, use the Scheduled Tasks MCP:
Tool: create_scheduled_task
- Schedule: 30 minutes before the meeting
- Action: Run meeting-prep skill for the next calendar eventThis is optional — only set up if the user requests automated prep.
| Scenario | Action | |----------|--------| | No upcoming meetings found | Report "No meetings found in the next 24 hours" and ask for a specific event | | Calendar MCP unavailable | Ask user for meeting details manually (title, attendees, topic) | | Clay enrichment fails | Skip company data, note "Enrichment unavailable" in attendee row | | Slack search returns nothing | Note "No recent Slack activity found" in context section | | Gmail search returns nothing | Note "No recent email threads found" in context section | | Attendee not in Slack | Mark as external, skip Slack enrichment |
~/Reports/ (local filesystem)| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→fail | 8,005 | 13,847 | +73% | 1 | 1 | 0% | 1,327 | 2,108 | +59% | 0 | 0 | — |
case-01 | fail→fail | 5,859 | 7,052 | +20% | 1 | 1 | 0% | 1,025 | 2,062 | +101% | 0 | 0 | — |
case-02 | fail→fail | 6,045 | 5,166 | -15% | 1 | 1 | 0% | 983 | 1,854 | +89% | 0 | 0 | — |
case-04 | fail→pass | 5,710 | 1,878 | -67% | 1 | 1 | 0% | 992 | 1,914 | +93% | 0 | 0 | — |
case-05 | pass→pass | 7,427 | 3,533 | -52% | 1 | 1 | 0% | 1,212 | 2,173 | +79% | 0 | 0 | — |
case-06 | fail→fail | 8,054 | 1,798 | -78% | 1 | 1 | 0% | 1,358 | 1,789 | +32% | 0 | 0 | — |
case-07 | fail→pass | 6,664 | 1,890 | -72% | 1 | 1 | 0% | 1,185 | 1,844 | +56% | 0 | 0 | — |
case-08 | fail→fail | 6,538 | 1,356 | -79% | 1 | 1 | 0% | 1,049 | 1,734 | +65% | 0 | 0 | — |
case-09 | fail→pass | 10,787 | 2,892 | -73% | 1 | 1 | 0% | 1,753 | 1,948 | +11% | 0 | 0 | — |
case-10 | fail→pass | 11,116 | 1,993 | -82% | 1 | 1 | 0% | 1,917 | 1,861 | -3% | 0 | 0 | — |
case-11 | fail→fail | 10,811 | 2,696 | -75% | 1 | 1 | 0% | 1,954 | 1,900 | -3% | 0 | 0 | — |
case-12 | fail→fail | 6,314 | 2,119 | -66% | 1 | 1 | 0% | 1,158 | 1,853 | +60% | 0 | 0 | — |
case-13 | fail→fail | 9,598 | 14,403 | +50% | 1 | 1 | 0% | 1,594 | 1,986 | +25% | 0 | 0 | — |
case-14 | fail→fail | 3,462 | 1,237 | -64% | 1 | 1 | 0% | 609 | 1,704 | +180% | 0 | 0 | — |
case-15 | fail→pass | 8,807 | 1,596 | -82% | 1 | 1 | 0% | 1,574 | 1,729 | +10% | 0 | 0 | — |
case-16 | fail→pass | 9,480 | 2,913 | -69% | 1 | 1 | 0% | 1,536 | 1,996 | +30% | 0 | 0 | — |
case-17 | fail→pass | 10,666 | 1,542 | -86% | 1 | 1 | 0% | 1,715 | 1,740 | +1% | 0 | 0 | — |
case-18 | fail→pass | 9,939 | 2,502 | -75% | 1 | 1 | 0% | 1,718 | 1,980 | +15% | 0 | 0 | — |
case-19 | pass→pass | 3,326 | 1,449 | -56% | 1 | 1 | 0% | 540 | 1,779 | +229% | 0 | 0 | — |
case-20 | pass→pass | 5,845 | 9,572 | +64% | 1 | 1 | 0% | 1,043 | 3,136 | +201% | 0 | 0 | — |
case-21 | fail→pass | 5,078 | 15,407 | +203% | 1 | 1 | 0% | 922 | 3,235 | +251% | 0 | 0 | — |
case-22 | fail→fail | 6,074 | 18,440 | +204% | 1 | 1 | 0% | 848 | 4,094 | +383% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 19 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +41 percentage points is the difference between those two pass rates over the 19 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/3/2026 | +18% |
Other measured skills in the registry, with their headline benchmark lift.