agentleFS
allthriveaiallthriveai/amplify/.claude/skills/meeting/SKILL.md

meeting

Skill0 starsChanged 3 months agoMITView on GitHub

Processes a Plaud-synced transcript or pasted meeting notes into a structured meeting note with decisions, action items, and attendees. Use when the user runs /meeting or wants to process a meeting recording.

14 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use commit and handoff.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • SentrySkill

    commit

    Create commit messages following Sentry conventions. Use when committing code changes, writing commit messages, or formatting git history. Follows conventional commits with Sentry-specific issue references.

    Also in 3 projects that use this

  • MicrosoftSkill

    handoff

    Compact the current conversation into a handoff document for another agent to pick up.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

12 rules. 0 are common, 3 are distinctive, and 9 are very specific to this project.

  • 0

    Common

    Most of the 13 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 9

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    When the user runs /meeting, optionally followed by a filename, Plaud note title, or pasted transcript:

    line 10
  2. Distinctive:

    Read the config and resolve the vault path.

    line 20
  3. Very specific:

    Read {vaultPath}/{paths.voice} (Voice.md) if it exists — shapes how you write summaries.

    line 22
  4. Very specific:

    Read {vaultPath}/{paths.voice} (Voice.md) if it exists — helps flag which decisions are worth ingesting into the wiki later.

    line 23
  5. Very specific:

    Read all existing meeting notes from {vaultPath}/{paths.meetings} (default: Sources/Meetings/).

    line 37
  6. Very specific:

    Avoid duplicates (check plaud_file_id in frontmatter)

    line 38
  7. Very specific:

    If a meeting with the same plaud_file_id already exists and is already processed, tell the user and ask if they want to re-process it.

    line 42
  8. Very specific:

    Create the filename: {date}-{slug}.md where {slug} is a short kebab-case title derived from the meeting topic (e.g., …

    line 75
  9. Distinctive:

    No AI vocabulary. No filler.

    line 98
  10. Very specific:

    Write the meeting note to {vaultPath}/{paths.meetings}/{filename}.

    line 102
  11. Very specific:

    Use the writeMeeting() function from src/vault/writer.ts if running programmatically, or write the file directly.

    line 104
  12. Distinctive:

    Keep it brief. Don't repeat the full content — they can read the note.

    line 146

“Common” is measured: kept by at least 6 in 10 of the 13 edited versions of this file we could read. “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

Meeting

Instructions

When the user runs /meeting, optionally followed by a filename, Plaud note title, or pasted transcript:

Step 0: Load Context

Find the .amplifyrc config file. Check these locations in order:

  1. .amplifyrc in the current working directory
  2. .amplifyrc at the path specified by VAULT_PATH environment variable
  3. .amplifyrc at the fallback vault path (if configured in CLAUDE.md or known from previous sessions)

Read the config and resolve the vault path.

Read {vaultPath}/{paths.voice} (Voice.md) if it exists — shapes how you write summaries. Read {vaultPath}/{paths.voice} (Voice.md) if it exists — helps flag which decisions are worth ingesting into the wiki later.

Step 1: Find the Source

Three input modes:

A. Plaud note reference — The user names a specific Plaud-synced note. Look in the Plaud sync folder (typically Plaud/ in the vault root — check for a folder matching this or similar names like plaud, Plaud Notes). Read the file, extract file_id from its frontmatter if present.

B. Unprocessed Plaud notes — If the user runs /meeting with no argument, scan the Plaud sync folder for notes that haven't been processed yet. List them and ask which one to process. If there's only one unprocessed note, confirm and proceed.

C. Pasted transcript — The user pastes transcript text directly. Use this as the source content. Set source: "manual".

Step 2: Read Existing Meetings

Read all existing meeting notes from {vaultPath}/{paths.meetings} (default: Sources/Meetings/). This helps:

  • Avoid duplicates (check plaud_file_id in frontmatter)
  • Maintain consistent formatting
  • Link to related meetings

If a meeting with the same plaud_file_id already exists and is already processed, tell the user and ask if they want to re-process it.

Step 3: Extract and Structure

From the raw transcript/summary, extract and organize into these sections:

Summary — 3-5 bullet points capturing what the meeting was about. Lead with outcomes, not process. "Decided to ship X by Friday" not "Discussed the timeline for X."

Key Decisions — Bulleted list of decisions made. Each decision should be a clear statement. If no explicit decisions were made, say so — don't fabricate them.

Action Items — Checkbox list with owners and due dates where mentioned:

- [ ] @PersonName: Do the thing (by Friday)
- [ ] @PersonName: Follow up on X

Extract these from commitments made during the meeting. If someone said "I'll do X" that's an action item.

Discussion Topics — Organized by topic, brief notes on what was discussed. Include dissenting views and unresolved questions.

Attendees — Extract from speaker labels in the transcript if available. If the Plaud note has speaker labels like "Speaker 1:", "Speaker 2:", list them. If you can infer names from context, use them.

Raw Transcript — Include the full transcript in a collapsible section:

<details>
<summary>Full Transcript</summary>

[transcript here]

</details>

Step 4: Build the Meeting Note

Create the filename: {date}-{slug}.md where {slug} is a short kebab-case title derived from the meeting topic (e.g., 2026-04-23-weekly-standup.md).

Build frontmatter:

---
title: "Meeting Title"
date: YYYY-MM-DD
duration: "Xm"
attendees:
  - Person 1
  - Person 2
source: plaud        # or "manual"
plaud_file_id: "abc" # only if from Plaud
tags:
  - meeting
processed: true
---

Step 5: Humanize

Apply humanizer rules to your written sections (Summary, Key Decisions, Discussion Topics). Do NOT modify the raw transcript or action items — those should reflect what was actually said.

No AI vocabulary. No filler. No significance inflation. Be direct.

Step 6: Write to Vault

Write the meeting note to {vaultPath}/{paths.meetings}/{filename}.

Use the writeMeeting() function from src/vault/writer.ts if running programmatically, or write the file directly.

Step 7: Emit Signal

Emit a meeting_synced signal:

{
  "type": "meeting_synced",
  "data": {
    "title": "Meeting Title",
    "date": "2026-04-23",
    "source": "plaud",
    "plaudFileId": "abc123",
    "path": "Meetings/2026-04-23-weekly-standup.md"
  }
}

Step 8: Offer the wiki handoff

The meeting note is a raw source. It lands in {paths.meetings} under the source layer, immutable from here on. On its own it is inert — a meeting nobody links to is a meeting nobody finds again.

Offer to ingest it: run the ingest skill against the note so the people mentioned become entity pages, recurring topics become concept pages, and the decisions get cross-referenced against what is already in the wiki.

Worth pushing on when the meeting named a decision, a person not yet in Wiki/Entities, or a topic that has come up in earlier meetings. Skip it for routine status calls — not every meeting is knowledge, and a wiki full of standups is a wiki nobody reads.

Step 9: Report

Tell the user:

  • Meeting title and date
  • Number of decisions and action items extracted
  • Path to the meeting note
  • Any connections to goals or existing meetings
  • Whether it was ingested into the wiki, and which pages it touched

Keep it brief. Don't repeat the full content — they can read the note.

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.