Update Todos from Conversation

Analyze the current conversation to identify incomplete work and update todo.json.

Process

1. •

   Read todo.json at project root (create if doesn't exist)

2. •

   Scan conversation for:

   • •Tasks marked as incomplete, blocked, or deferred
   • •Errors encountered that need follow-up
   • •Features partially implemented
   • •"TODO", "FIXME", "later", "next time" mentions
   • •Explicit user requests not yet fulfilled
3. •

   Auto-archive: match completed work against active tasks

4. •

   Propose changes to user:

   • •Tasks to archive (with reason)
   • •New tasks to add (with suggested priority)
   • •Never add tasks without asking
5. •

   Write todo.json after user approval

6. •

   Display active tasks sorted by priority

7. •

   Output session summary (for searchability):

═══════════════════════════════════════════════════
SESSION_TAG_<topic-slug>_<YYYYMMDD>
Summary: <30 words max, main topics covered>
Todos: <count active> active, <count archived this session> archived
═══════════════════════════════════════════════════

File Format

{
  "active": [
    {
      "id": "t1",
      "task": "Implement OAuth2 flow",
      "context": "Need Google + GitHub providers",
      "priority": "high",
      "created": "2026-01-26",
      "tags": ["auth", "feature"]
    }
  ],
  "archived": [
    {
      "id": "t0",
      "task": "Setup project structure",
      "completed": "2026-01-26",
      "tags": ["setup"]
    }
  ]
}

Fields

Rules

• •Never invent tasks — only track what was actually discussed
• •IDs are never reused — always increment from max existing
• •Don't delete tasks — archive them with completion date
• •Prune archive if >50 items (keep most recent)
• •Add todo.json to .gitignore if not already there