When to Use This Skill

/meeting-notes

Create structured meeting notes from raw input (transcript, bullet points, or audio summary). Uses three parallel post-processing agents to extract decisions, action items, and topic summaries from the captured content.

When to Use This Skill

• Recording meeting outcomes from notes or transcripts
• Structuring raw meeting notes into a consistent format
• Extracting decisions and action items from meeting content
• Creating searchable, linkable meeting records

Usage

/meeting-notes "<meeting-title>" [--date YYYY-MM-DD] [--attendees "Name1, Name2"]

Parameters

Instructions

Phase 1: Capture Raw Content

Prompt the user for meeting content in any format:

• Transcript — Paste a meeting transcript
• Bullet points — Quick notes taken during the meeting
• Audio summary — Summary from a voice note or recording
• Mixed — Any combination of the above

Also gather:

• Attendees — Who was in the meeting
• Context — Related project or workstream
• Agenda — What was planned to discuss (if available)

Phase 2: Post-Processing — Agent Team

After the raw content is captured, launch three agents simultaneously using the Task tool.

Agent 1: Decision Extractor (Haiku) Task: Find all decisions made during the meeting

• Scan content for decision language ("decided", "agreed", "approved", "will go with")
• For each decision: what was decided, who decided, context, and implications
• Categorise: formal (needs ADR), informal (team agreement), action (task assignment)
• Flag decisions that need follow-up documentation Return: List of { decision, deciders, category, needsFollowUp }

Agent 2: Action Item Extractor (Haiku) Task: Extract all action items and assignments

• Scan content for action language ("will do", "action:", "to-do", "by next week", names followed by verbs)
• For each action: what needs to be done, who owns it, deadline (if mentioned)
• Flag actions without clear owners
• Flag actions without deadlines Return: List of { action, owner, deadline, status }

Agent 3: Topic Summariser (Sonnet) Task: Summarise each discussion topic

• Identify distinct topics discussed in the meeting
• For each topic: summarise the discussion in 2-3 sentences
• Note any unresolved questions or parking lot items
• Identify topics that need follow-up in future meetings
• Link to relevant projects, systems, or concepts mentioned Return: List of { topic, summary, status, relatedNotes[] }

Phase 3: Generate Meeting Note

Combine raw content and agent outputs into a structured meeting note.

Output Format

---
type: Meeting
title: "Meeting - YYYY-MM-DD <Meeting Title>"
created: YYYY-MM-DD
date: YYYY-MM-DD
attendees:
  - "[[Person Name]]"
project: "[[Project - Name]]"
tags: [activity/meeting, project/relevant-tag]
summary: "<One-line summary of meeting outcome>"
---

# Meeting - YYYY-MM-DD <Meeting Title>

## Attendees

- [[Person 1]]
- [[Person 2]]

## Agenda

1. <Topic 1>
2. <Topic 2>
3. <Topic 3>

## Discussion

### <Topic 1>

<2-3 sentence summary of discussion>

- Key point discussed
- Alternative considered
- Conclusion reached

### <Topic 2>

<Summary>

## Decisions

| # | Decision                          | Deciders        | Follow-up Needed |
|---|-----------------------------------|-----------------|------------------|
| 1 | <What was decided>                | <Who decided>   | Yes — Create ADR |
| 2 | <What was decided>                | <Who decided>   | No               |

## Action Items

| # | Action                            | Owner           | Deadline    | Status  |
|---|-----------------------------------|-----------------|-------------|---------|
| 1 | <What needs to be done>           | [[Person]]      | YYYY-MM-DD  | Pending |
| 2 | <What needs to be done>           | [[Person]]      | YYYY-MM-DD  | Pending |

## Parking Lot

- <Topics deferred to future meetings>

## Related Notes

- [[Project - Relevant Project]]
- [[ADR - Relevant Decision]]

Examples

Example 1: From Transcript

/meeting-notes "Sprint Planning" --date 2026-02-10 --attendees "Alice, Bob, Charlie"

User pastes transcript; agents extract decisions, actions, and topic summaries.

Example 2: From Bullet Points

/meeting-notes "Architecture Review"

User provides bullet-point notes; agents structure them into a full meeting record.

Example 3: Quick Capture

/meeting-notes "1:1 with Manager" --attendees "Manager Name"

Minimal input for a quick meeting capture with decision and action extraction.

Invoke with: /meeting-notes "<title>" then paste or type the meeting content