Meeting Recap

Meeting Recap Skill

Purpose

After a meeting, quickly extract and organize everything important from the transcript — decisions made, action items assigned, technical details discussed, open questions, and follow-ups. Designed for highly technical engineering meetings where precision matters and no detail should be lost.

Required Configuration

This skill requires the following from the user's global copilot instructions (~/.copilot/copilot-instructions.md):

• ## VIP Contacts — for flagging action items and contributions from key people
• ## Meeting Keywords (optional) — recurring meeting keyword strategies for better topic extraction
• ## Notes Repo (optional) — path to personal notes vault, for saving the recap

If any required configuration is missing, ask the user for the values and offer to add them to their ~/.copilot/copilot-instructions.md.

How to Execute

Step 1: Identify the Meeting

Ask the user for the meeting name if not provided. Then fetch meeting details:

workiq ask -q "Find the most recent meeting titled '{meeting title}'. Return the date, time, duration, attendees, and full transcript."

If multiple matches are found, present them and ask the user to confirm which one.

If the transcript is truncated or incomplete, run a follow-up query:

workiq ask -q "Continue returning the transcript for the meeting '{meeting title}' on {date}. Pick up where you left off."

Repeat until the full transcript is retrieved.

Step 2: Analyze the Transcript

With the full transcript in hand, perform a thorough analysis covering every category below. Do NOT summarize loosely — extract specific details, exact quotes where relevant, and attribute statements to speakers.

Analysis categories:

1. Meeting overview — Date, duration, attendees, and the general purpose/agenda of the meeting.

2. Key topics discussed — Every distinct topic or agenda item covered. For each topic:

   • What was discussed
   • Who drove the discussion
   • What conclusions were reached (if any)

3. Technical details — Architecture decisions, design patterns, code changes, system behaviors, performance metrics, infrastructure changes, API contracts, data models, or any other technical specifics mentioned. Be precise — include version numbers, service names, config values, error codes, metric thresholds, and any concrete technical artifacts referenced.

4. Decisions made — Explicit decisions the group committed to. Include:

   • The decision itself
   • Who made or approved it
   • Any conditions or caveats

5. Action items — Tasks assigned or volunteered for. For each:

   • What needs to be done
   • Who owns it
   • Any stated deadline or urgency
   • Flag items owned by VIP contacts (from ## VIP Contacts in global instructions)

6. Open questions & unresolved items — Topics raised but not resolved, questions left unanswered, or items explicitly deferred.

7. Risks & concerns raised — Any risks, blockers, or concerns mentioned by attendees.

8. Follow-ups & next steps — Agreed-upon next steps, follow-up meetings, or future checkpoints.

Step 3: Present the Recap

Format the output as follows:

═══════════════════════════════════════════════════
  📝 MEETING RECAP — {meeting title}
  📅 {date} | ⏱️ {duration} | 👥 {attendee count} attendees
═══════════════════════════════════════════════════

👥 ATTENDEES
  • {name1}, {name2}, {name3}, ...

🎯 MEETING OVERVIEW
  {1-3 sentence summary of the meeting's purpose and outcome}

💡 KEY TOPICS DISCUSSED
  1. {Topic Title}
     • {details, who drove it, conclusions}
  2. {Topic Title}
     • {details}
  ...

🔧 TECHNICAL DETAILS
  • {Specific technical detail with exact values}
  • {Architecture or design decision with rationale}
  • {Code change, API contract, or system behavior discussed}
  ...

✅ DECISIONS MADE
  • {Decision} — decided by {who}, {any caveats}
  ...

📋 ACTION ITEMS
  • [ ] {Task description} — Owner: {name} {⭐ if VIP} {deadline if stated}
  • [ ] {Task description} — Owner: {name}
  ...

❓ OPEN QUESTIONS
  • {Question or unresolved item} — raised by {who}
  ...

⚠️ RISKS & CONCERNS
  • {Risk or blocker} — raised by {who}
  ...

➡️ NEXT STEPS
  • {Follow-up action or checkpoint}
  ...

═══════════════════════════════════════════════════

Step 4: Offer Follow-up Actions

After presenting the recap, offer:

1. Save to notes — "Would you like me to save this recap to your notes vault?"
   • If yes, save to {NOTES_PATH}/domains/{meeting-domain}/ or {NOTES_PATH}/inbox/ using the notes skill
2. Create ADO work items — "Would you like me to create work items for any of the action items?"
3. Share via clipboard — "Would you like me to copy this recap to your clipboard for sharing?"

Constraints

• This is read-only by default — do not create work items, send messages, or modify anything unless the user explicitly opts in via the follow-up actions
• Do not fabricate or infer information not present in the transcript — if something is unclear, flag it as an open question
• Attribute statements and action items to specific people whenever the transcript makes this clear
• Preserve technical precision — do not simplify or generalize technical details (keep exact names, numbers, versions, and config values)
• If the transcript is unavailable or empty, inform the user and suggest they check whether the meeting was recorded
• Keep the recap comprehensive but scannable — use bullet points, not paragraphs
• Omit any section that has no content (e.g., if no risks were raised, skip the Risks section)

Examples

Basic usage

User: "Recap the AET Hub Hour meeting"
→ Skill fetches transcript via workiq, analyzes it, produces the structured recap

With date disambiguation

User: "Recap the Triitus sync"
→ workiq returns multiple matches
→ Skill presents: "I found 3 instances of Triitus Weekly Sync. Which one?"
   1. Feb 25, 2026 (most recent)
   2. Feb 18, 2026
   3. Feb 11, 2026
→ User picks one, skill proceeds with that transcript