Scheduling

Most agent workflows are time-triggered. Three approaches:

Approach	Pros	Cons
LaunchAgents (macOS)	Reliable, runs without Claude	No reasoning, just scripts
Claude Code cron commands	Can reason about edge cases	Needs active session
Hybrid (recommended)	Scripts collect data, agents reason	More complex but most capable

Recommended: Hybrid Model

Scripts run on schedule via LaunchAgents (data collection, simple routing)
Agent reasoning triggered when classification/drafting is needed
Results pushed via SMS/email/Telegram/other channels

LaunchAgents (macOS Scheduled Tasks)

LaunchAgents are .plist files in ~/Library/LaunchAgents/ that run scripts on schedule.

Example LaunchAgent Setup

Agent	Schedule	Purpose
`com.yourname.nightcrew`	3:00 AM daily	Nightly data pipeline
`com.yourname.morningbrief`	6:00 AM daily	Morning digest generation
`com.yourname.todo-sync`	Every 10 min	Sync todos with a note-taking app
`com.yourname.n8n`	At startup	n8n workflow engine

Example: Nightly Pipeline

An example nightly automation pipeline that collects data across sources:

Stage	What It Does
1	File triage (Downloads -> PARA folders)
2	Claude Code session context extraction
3	Email digest + cleanup
4	Message extraction (WhatsApp, iMessage, etc.)
5	Voice memo transcription
6	Calendar events for today/tomorrow
7	Photo triage and classification
8	Activity summary compilation
9	Morning brief generation
10	Git push to sync repo

The key pattern is: collect -> aggregate -> output to vault -> push.


  ┌─ Schedule (LaunchAgents) ───────────────────────────────────┐
  │                                                             │
  │  3 AM NightCrew ──▶ 6 AM Morning Brief ──▶ Every 10m        │
  │                                             Todo Sync       │
  └────────┬────────────────────────────────────────────────────┘
           │
           ▼
  ┌─ Pipeline Stages ───────────────────────────────────────────┐
  │                                                             │
  │  File Triage ──▶ Data Extract ──▶ Aggregate ──▶ Output      │
  │                                                  │          │
  │                                                  ▼          │
  │                                             Push to Vault   │
  │                                                             │
  └─────────────────────────────────────────────────────────────┘

Setting Up a New LaunchAgent


<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN"
  "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
    <key>Label</key>
    <string>com.yourname.example-task</string>
    <key>ProgramArguments</key>
    <array>
        <string>/usr/bin/python3</string>
        <string>/path/to/script.py</string>
    </array>
    <key>StartCalendarInterval</key>
    <dict>
        <key>Hour</key>
        <integer>8</integer>
        <key>Minute</key>
        <integer>0</integer>
    </dict>
    <key>EnvironmentVariables</key>
    <dict>
        <key>OP_DISABLED</key>
        <string>1</string>
        <key>PATH</key>
        <string>/usr/local/bin:/usr/bin:/bin</string>
    </dict>
    <key>StandardOutPath</key>
    <string>/tmp/example-task.log</string>
    <key>StandardErrorPath</key>
    <string>/tmp/example-task-error.log</string>
</dict>
</plist>

Important: LaunchAgents cannot use 1Password CLI (no GUI session for biometric). Set OP_DISABLED=1 and use .env files as fallback.


# Load a LaunchAgent
launchctl load ~/Library/LaunchAgents/com.yourname.example-task.plist
 
# Unload
launchctl unload ~/Library/LaunchAgents/com.yourname.example-task.plist
 
# Check status
launchctl list | grep com.yourname