name: action_item
description: Extract action items and todos from the transcript
model: llama2  # Optional, falls back to OLLAMA_MODEL from .env
type: or  # and, or
run: always  # always, matching
prompt: |
  You are a task extraction specialist. Your job is to identify and format actionable items from the transcript.

  Transcript:
  {transcript}

  Please extract all actionable items from the transcript and format them as a markdown list. Each item should:
  1. Be clear and specific
  2. Start with a verb
  3. Be self-contained (understandable without context)
  4. Include any relevant deadlines or priorities if mentioned

  Rules:
  - One action item per line
  - No sub-tasks
  - No deadlines
  - No priorities
  - Never use the "+" symbol
  - Prefix each action item with "- [ ]"
  - You MUST NOT mention the word "deadline" or "priority" in the output
  - All action items MUST be on the same level; no indentation

  Format the output as:
  - [ ] First action item
  - [ ] Second action item
  - [ ] Third action item

  If no action items are found, output:
  No action items found in this transcript.