agentleFS
JustinTSmithJustinTSmith/ai-operator-skills/meeting-intel/SKILL.md

meeting-intel

Skill0 starsChanged 4 months agoMITView on GitHub

Extract decisions, action items, and open questions from a meeting transcript, then create tasks in the user's tracker. Use when the user pastes or points to a meeting transcript, recording summary, or call notes, or says "process this meeting", "what were the action items", or "turn this call into tasks".

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read the full transcript before extracting anything.

    line 22
  2. Distinctive:

    Create the action items in the user's tracker, in this order of preference: a connected task MCP tool (Linear, Asana, Todoist, etc.) if …

    line 35
  3. Distinctive:

    When creating tracker tasks: title is the action, description carries the transcript quote that generated it, owner and due date set as …

    line 39
  4. Distinctive:

    Do not summarize the meeting.

    line 61
  5. Distinctive:

    Do not soften ownership: "the team will look into it" becomes a named owner or a flagged item, never a shared vague one.

    line 62
  6. Distinctive:

    Do not invent action items from suggestions nobody agreed to.

    line 64

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Meeting Intel

Turn a raw meeting transcript into structured, owned, dated follow-through.

Distilled from a pipeline that polls Fathom for recordings, extracts action items with Claude, and creates Linear tasks automatically. This skill is the extraction and task-creation core, made source-agnostic.

Inputs

Accept any of: pasted transcript text, a file path, a Fathom/Zoom/Meet summary, or rough typed notes. If the input is missing attendee names, ask once for them; otherwise never block on questions.

Workflow

  1. Read the full transcript before extracting anything.
  2. Extract into three buckets, strictly separated:
    • Decisions - things that were settled. Past tense, one line each, with who made the call if identifiable.
    • Action items - each MUST have an owner and a due date. If the transcript states neither, infer the owner from context and mark the date (proposed) with your best inference. Never emit an ownerless action item; assign it to the user and flag it.
    • Open questions - raised but not resolved. These are the follow-up agenda for next time.
  3. Discard filler: status updates with no consequence, small talk, scheduling chatter. A 60-minute meeting typically yields 3-8 action items. If you extracted 20, you are transcribing, not extracting.
  4. Create the action items in the user's tracker, in this order of preference: a connected task MCP tool (Linear, Asana, Todoist, etc.) if available in the session; otherwise the project's TODO/tasks file if one exists; otherwise output a markdown task block.
  5. When creating tracker tasks: title is the action, description carries the transcript quote that generated it, owner and due date set as extracted. Never create duplicates - search the tracker for similar open tasks first.

Output format

## Meeting: <title or best inference> - <date>

### Decisions
- <decision> (<who>)

### Action items          [created in <tracker> / markdown only]
- [ ] <action> - @<owner>, due <date>

### Open questions
- <question>

Anti-patterns

  • Do not summarize the meeting. This skill extracts commitments, not vibes.
  • Do not soften ownership: "the team will look into it" becomes a named owner or a flagged item, never a shared vague one.
  • Do not invent action items from suggestions nobody agreed to.

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.