agentleFS
timominkustimominkus/obsidian-para-gtd-ai/.claude/skills/process-meeting/SKILL.md

process-meeting

Skill9 starsChanged 8 months agoMITView on GitHub

Processes a meeting into a structured Obsidian note (YYYY-MM-DD_Meeting-Name.md). Use this skill whenever the user wants to capture, process, or document a meeting — whether they have a transcript (docx/txt/vtt file) or are working from memory. Trigger on phrases like "process meeting", "document meeting", "meeting notes", "process transcript", or any request to turn meeting content into a vault note. Always use this skill when a meeting is involved — don't try to do it ad-hoc without the skill.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

25 rules. About 0 are common, 21 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈21

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before anything else, determine which meeting is to be processed.

    line 23
  2. Distinctive:

    Look for a meeting name mentioned in notes or any explicit reference to a meeting today.

    line 26
  3. Distinctive:

    If multiple candidates are found, list them briefly and ask the user which one to process.

    line 34
  4. Distinctive:

    If none are found, ask the user for the meeting name / date before continuing.

    line 35
  5. Distinctive:

    After identifying the target folder, check whether a CLAUDE.md exists in that folder.

    line 44
  6. Distinctive:

    If found, read it and follow its instructions throughout the workflow.

    line 45
  7. Distinctive:

    After identifying the meeting, look for a transcript in this order:

    line 51
  8. Distinctive:

    Ask the user to confirm the correct file if multiple match.

    line 59
  9. Distinctive:

    If the meeting note already has an Agenda, use it as the skeleton.

    line 79
  10. Distinctive:

    If no agenda exists, derive topics from the transcript based on natural topic shifts.

    line 80
  11. Distinctive:

    Write what was said/concluded, not who said it.

    line 103
  12. Distinctive:

    Use the brain dump, INBOX note, and any content already in the prepared meeting note.

    line 110
  13. Distinctive:

    Wait for confirmation. Apply all corrections before proceeding.

    line 149
  14. Distinctive:

    If no prepared meeting note was found: create a new file using the structure below.

    line 156
  15. Very specific:

    Write detailed notes per agenda item as nested bullet points.

    line 162
  16. Distinctive:

    Add only if decisions were explicitly made.

    line 199
  17. Very specific:

    After supplementing the meeting note, update relevant hub files in 03_AREAS/People/.

    line 205
  18. Distinctive:

    Only update sections that are relevant.

    line 216
  19. Very specific:

    Read the Daily Note template first (04_RESOURCES/Templates/Daily Note.md) to match section names.

    line 223
  20. Distinctive:

    Add 2–4 bullet points with the most important takeaways

    line 227
  21. Distinctive:

    Use the Section Anchoring Rule: anchor edits at the target section's own header plus the last existing item — never at the next section's …

    line 233
  22. Distinctive:

    After all content has been transferred and confirmed:

    line 239
  23. Distinctive:

    Use the meeting date

    line 244
  24. Distinctive:

    When creating a meeting note from scratch:

    line 256
  25. Very specific:

    All internal references use [[WikiLinks]] — never backtick-wrapped

    line 317

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Process Meeting Skill

Processes a meeting and supplements an existing meeting note in the vault (or creates one). Identifies the meeting, loads available inputs (brain dump, prepared note, transcript), enriches the note with structured content, updates hub files, syncs the daily note, and cleans up source files.


Step 0 — Identify the meeting

Before anything else, determine which meeting is to be processed. Search in this order:

  1. Daily note brain dump — Read today's daily note (01_ACTIVE/Daily Notes/YYYY-MM-DD.md). Look for a meeting name mentioned in notes or any explicit reference to a meeting today.
  2. INBOX note — Glob 00_INBOX/**/*.md and scan for a note that looks like a meeting capture (title contains a date or person's name, or contains an agenda list).
  3. Existing meeting note — Glob 03_AREAS/**/*.md for a file dated today or recently that looks like a prepared meeting note (has an Agenda or Attendees field).

If multiple candidates are found, list them briefly and ask the user which one to process. If none are found, ask the user for the meeting name / date before continuing.

Once identified, note:

  • Meeting type (1:1, team meeting, external, etc.)
  • Participants
  • Date
  • File path (this is the file that will be supplemented, not replaced)

Check for meeting-specific instructions: After identifying the target folder, check whether a CLAUDE.md exists in that folder. If found, read it and follow its instructions throughout the workflow.


Step 1 — Find and read the transcript

After identifying the meeting, look for a transcript in this order:

  1. Mentioned in the prompt — the user explicitly named a file path. Use it directly.
  2. Uploaded in this chat — check for any recently added .docx, .txt, or .vtt file.
  3. Downloads folder — search the user's Downloads folder for a .docx, .txt, or .vtt file whose name matches the meeting (by participant name, date, or topic keywords). Ask the user to confirm the correct file if multiple match.
  4. No transcript available — proceed in memory-only mode (Step 2B).

Read the transcript:

  • For .txt / .vtt: Read tool directly. For .vtt: strip header and timing lines; merge consecutive lines from the same speaker.
  • For .docx: Extract with python-docx:
    python3 -c "
    import docx, sys
    doc = docx.Document(sys.argv[1])
    print('\n'.join(p.text for p in doc.paragraphs))
    " /path/to/file.docx

Step 2A — Extract content (with transcript)

Agenda structure:

  • If the meeting note already has an Agenda, use it as the skeleton. Map transcript content onto topics.
  • If no agenda exists, derive topics from the transcript based on natural topic shifts.
  • Mark any agenda item not covered as *(not discussed)*.

Decisions — ONLY include this section if an explicit decision was made (someone said "we decided", "we agreed to", etc.). Implicit conclusions do not count.

Tasks — STRICT RULE: only capture tasks where a specific person verbally commits in the transcript.

CategoryWhen to use
My Tasks — Today 🔴You commit to something due today or urgent
My Tasks — This Week 📅You commit to something for this week
My Tasks — Next Week 📅You commit to something for next week
Waiting For ⌛Someone else commits to delivering something you need
Delegated 👥You explicitly ask someone to own a task and they accept
Agenda Items 📌You say you need to raise a topic in a different meeting

What does NOT count:

  • Logical next steps not spoken aloud
  • Topics discussed without a named owner
  • Exploratory ideas or suggestions without acceptance
  • Anything you have to infer rather than hear

Notes — detailed, factual summary per agenda item as nested bullet points.

  • Write what was said/concluded, not who said it.
  • Attribution in parentheses is acceptable when it adds meaningful context. Keep it rare.

Step 2B — Extract content (from memory only)

Use the brain dump, INBOX note, and any content already in the prepared meeting note. Same extraction logic as 2A — decisions only if explicit, tasks only if committed.


Step 3 — Preview in chat — REQUIRED before writing anything

Present a structured preview:

## Preview: [Meeting Name] — YYYY-MM-DD

### Agenda
- Topic A
  - Sub-topic A1
- Topic B *(not discussed)*

### Decisions
*(none)* — or list them

### Tasks
**My Tasks — Today**
- [ ] ...

**My Tasks — This Week**
- [ ] ...

**Waiting For**
- [ ] ⌛ Name | ...

**Delegated**
- [ ] 👥 Name | ...

**Agenda Items**
- [ ] 📌 @Name | ...

Then say: "Does this look right? I'll then update the meeting note, hub files, and daily note."

Wait for confirmation. Apply all corrections before proceeding.


Step 4 — Supplement or create the meeting note

If a prepared meeting note was found in Step 0: supplement it in place. If no prepared meeting note was found: create a new file using the structure below.

Fill in or replace these sections:

4a · Notes section

Write detailed notes per agenda item as nested bullet points. Structure: top-level agenda items → ### Heading, sub-items → #### Heading.

4b · Tasks section

## Tasks

### 🔴 My Tasks — Today

- [ ] 🔴 | Topic | Description | [[meeting-link]] | YYYY-MM-DD

### 📅 My Tasks — This Week

- [ ] 📅 | Topic | Description | [[meeting-link]] | YYYY-MM-DD

### 📅 My Tasks — Next Week

- [ ] 📅 | Topic | Description | [[meeting-link]] | YYYY-MM-DD

### ⏳ Waiting For

- [ ] ⌛ Name | Topic | Description | [[meeting-link]] | Due date

### 👥 Delegated

- [ ] 👥 Name | Topic | Description | [[meeting-link]] | Due date

### 📌 Agenda Items

- [ ] 📌 @Name | Topic | Source | [[meeting-link]]

Omit any sub-section that has no items. If no tasks at all: *No tasks.*

4c · Decisions section

Add only if decisions were explicitly made. Insert before Tasks.


Step 5 — Update hub files

After supplementing the meeting note, update relevant hub files in 03_AREAS/People/.

What to update:

  1. 📌 Next Meeting queue — Add topics to raise in the next session with this person. Format: - [ ] 📌 Topic | Context | [[current-meeting-link]]
  2. Tasks (👥 / ⌛) — Add any delegated or waiting-for items for this person.
  3. Session link — Add a link to the processed meeting note in the session log section.

Only update sections that are relevant. Do not restructure the hub file.


Step 6 — Update the daily note

Find today's daily note in 01_ACTIVE/Daily Notes/YYYY-MM-DD.md. Read the Daily Note template first (04_RESOURCES/Templates/Daily Note.md) to match section names.

Updates:

  1. Under ## Notes: add - Meeting processed: [[YYYY-MM-DD_Meeting-Name]]
  2. Add 2–4 bullet points with the most important takeaways
  3. Your own tasks → under ### 🔴 Heute or ### 📅 Diese Woche
  4. Waiting For → under ### ⏳ Waiting For
  5. Delegated → under ### 👥 Delegiert (or Hub file if long-running)
  6. Agenda Items → under ### 📌 Agenda Items

Use the Section Anchoring Rule: anchor edits at the target section's own header plus the last existing item — never at the next section's header.


Step 7 — Cleanup

After all content has been transferred and confirmed:

  1. Archive the transcript — move it to 06_ARCHIVE/YYYY/MM_Month/Transcripts/ with a standardized name. Never delete it.

    Archive filename format: YYYY-MM-DD_[Meeting-Title]_[Participants].[ext]

    • Use the meeting date
    • Short, descriptive meeting title (2–4 words, kebab-case)
    • Participants as dash-separated list (first names only)
  2. Archive the INBOX note — if a capture note was found in 00_INBOX/, move it to 06_ARCHIVE/YYYY/MM_Month/Notes/. Never delete it.

Confirm to the user: "Cleaned up: transcript archived, inbox note archived."


Meeting note structure reference

When creating a meeting note from scratch:

---
tags: [meeting-notes, YYYY-MM-DD]
up: "[[03_AREAS/Organization/Meeting-Type]]"
---

# YYYY-MM-DD — Meeting Name

**Date:** YYYY-MM-DD
**Attendees:** Name, Name, ...

---

## Agenda

- [[#Topic A|Topic A]]
- [[#Topic B|Topic B]]

---

## Decisions

*(omit if none)*

---

## Tasks

*(see Step 4b for structure)*

---

## Notes

### Topic A

- Bullet point

### Topic B

*Not discussed.*

File location routing:

Meeting typeFolder
1:1 with direct report03_AREAS/People/Team/[Name]/
1:1 with peer / colleague03_AREAS/People/Peers/[Name]/
1:1 with leadership03_AREAS/People/Leadership/[Name]/
Recurring team meeting03_AREAS/Organization/[MeetingType]/
General / external03_AREAS/Organization/

Quality checks before finishing

  • No invented tasks
  • Decisions section only present if explicitly flagged in source material
  • All internal references use [[WikiLinks]] — never backtick-wrapped
  • File name uses ISO date prefix: YYYY-MM-DD_
  • Notes written factually — no "X said that..." phrasing
  • Hub file updated with next-meeting topics and delegated tasks
  • Daily note updated with meeting link, key points, and your own tasks
  • Transcript archived with standardized name
  • Inbox note archived after content is transferred

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.