zhimeng-agent: Personal Automation Tasks

> Personal automation task collection using macOS launchd for scheduled execution of daily tasks with multi-platform synchronization.

Overview

A Python-based automation system that runs four scheduled tasks daily:

Project Structure

```

zhimeng-agent/

├── tasks/

│ ├── config.py # Unified configuration

│ ├── sync_utils.py # Multi-platform sync utilities

│ ├── daily_report/ # Daily report generator (0:00)

│ ├── email_organizer/ # Email organizer (2:00)

│ ├── desktop_organizer/ # Desktop organizer (4:00)

│ ├── tech_news/ # Tech news aggregator (7:00)

│ └── launchd/ # macOS launchd configurations

├── logs/ # Execution logs

└── .env # Environment variables

```

Task Implementations

1. Daily Report (`daily_report/`)

**Purpose**: Generate structured daily work reports from code commits and AI sessions

**Data Sources**:

**Implementation Steps**:

1. Scan configured Git repositories for commits within the last 24 hours

2. Extract commit messages, authors, and timestamps

3. Parse Claude Code session directories for interaction logs

4. Generate structured Markdown report with sections: Code Commits, AI Sessions, Statistics

5. Sync to platforms: Git, Notion, Feishu, Obsidian

**Output Format**:

```markdown

YYYY-MM-DD Work Report

Code Commits

Claude Code Sessions

Statistics

```

**Manual Execution**:

```bash

poetry run python -m tasks.daily_report.main --dry-run # Preview only

poetry run python -m tasks.daily_report.main # Generate and sync

```

2. Email Organizer (`email_organizer/`)

**Purpose**: Automatically archive notification emails using himalaya CLI

**Prerequisites**:

1. Install himalaya: `brew install himalaya`

2. Configure Gmail App Password

3. Create `~/.himalaya-config/himalaya/config.toml`

**Archive Rules** (prioritized):

**Implementation Steps**:

1. Connect to Gmail via himalaya CLI

2. Fetch unread emails from inbox

3. Apply rule-based classification (sender domain, subject patterns)

4. Move matched emails to Archive label

5. Generate statistics report by sender category

6. Sync summary to Obsidian

**Manual Execution**:

```bash

poetry run python -m tasks.email_organizer.main --dry-run # Analysis only

poetry run python -m tasks.email_organizer.main # Execute archiving

```

3. Desktop Organizer (`desktop_organizer/`)

**Purpose**: Clean up desktop and downloads folders by file type

**Organization Rules**:

| File Type | Target Directory |

|-----------|------------------|

| Images (.png, .jpg, .gif, .svg) | ~/Pictures/Desktop-Archive/ |

| Documents (.pdf, .doc, .md, .txt) | ~/Documents/Desktop-Archive/ |

| Archives (.zip, .tar, .gz, .rar) | ~/Downloads/Archives/ |

| Code (.py, .js, .java, .cpp, .go) | ~/Documents/Code-Archive/ |

| Other | ~/Documents/Desktop-Other/ |

**Safety Features**:

**Implementation Steps**:

1. Scan ~/Desktop and ~/Downloads

2. Identify files older than 7 days

3. Classify by extension using rule mapping

4. Create target directories if they don't exist

5. Move files to appropriate archive folders

6. Clean up files older than 30 days in download archives

7. Generate summary report with file counts and moved sizes

8. Sync log to Obsidian

**Manual Execution**:

```bash

poetry run python -m tasks.desktop_organizer.main --dry-run # Preview only

poetry run python -m tasks.desktop_organizer.main # Execute cleanup

```

4. Tech News Aggregator (`tech_news/`)

**Purpose**: Generate daily tech briefing from Hacker News and GitHub Trending

**Data Sources**:

**Output Format**:

```markdown

YYYY-MM-DD (Weekday) Tech Briefing

Hacker News Top Stories

1. [Article Title](url) (320 points, 45 comments)

GitHub Trending

> Repository description

```

**Implementation Steps**:

1. Fetch Hacker News top stories via official API

2. Scrape GitHub Trending page for top repositories

3. Extract metadata: points, comments, stars, languages, descriptions

4. Generate formatted Markdown report

5. Sync to Obsidian (News/ folder) and Feishu

**Manual Execution**:

```bash

poetry run python -m tasks.tech_news.main --dry-run # Generate only

poetry run python -m tasks.tech_news.main --hn-count 20 # Custom counts

poetry run python -m tasks.tech_news.main # Generate and sync

```

Configuration

Environment Variables (`.env`)

```bash

Feishu (required for Feishu sync)

FEISHU_APP_ID=your_app_id

FEISHU_APP_SECRET=your_app_secret

OpenAI (optional, for AI-powered summaries)

OPENAI_API_KEY=your_api_key

Notion (optional)

NOTION_TOKEN=your_token

```

Core Configuration (`tasks/config.py`)

```python

@dataclass

class TaskConfig:

# Schedule times (hour of day, 24-hour format)

DAILY_REPORT_HOUR: int = 0

EMAIL_ORGANIZE_HOUR: int = 2

DESKTOP_ORGANIZE_HOUR: int = 4

TECH_NEWS_HOUR: int = 7

# Path configuration

IDEAS_ROOT: Path = Path("/Users/username/workspace/ideas")

OBSIDIAN_VAULT: Path = Path("/Users/username/Documents/Obsidian Vault")

# Feishu recipient

FEISHU_RECIPIENT_OPEN_ID: str = "ou_xxxx"

```

Multi-Platform Sync

Obsidian

Feishu (Lark)

Notion

Git

Installation & Management

Install Scheduled Tasks

```bash

cd /path/to/zhimeng-agent/tasks/launchd

./install.sh install

```

This creates and loads four launchd plist files:

Check Task Status

```bash

./install.sh status

```

Shows loaded status for all four tasks.

Trigger Task Manually

```bash

./install.sh run daily-report

./install.sh run tech-news

```

Uninstall Tasks

```bash

./install.sh uninstall

```

Unloads and removes all launchd plist files.

View Logs

```bash

Standard output

tail -f logs/daily-report.log

tail -f logs/tech-news.log

Error logs

tail -f logs/daily-report.error.log

tail -f logs/email-organizer.error.log

```

Troubleshooting

Task Not Executing

```bash

Check if tasks are loaded

launchctl list | grep zhimeng

Manually trigger for testing

launchctl start com.zhimeng.tech-news

View system logs

log show --predicate 'subsystem == "com.apple.launchd"' --last 1h | grep zhimeng

```

Python Environment Issues

```bash

Verify Poetry virtualenv path

poetry env info --path

Test module imports

poetry run python -c "from tasks.config import config; print(config)"

```

Permission Issues

```bash

Ensure scripts are executable

chmod +x tasks/launchd/install.sh

Verify Python executable exists

ls -la $(poetry env info --path)/bin/python

```

Adding New Tasks

1. **Create task module**:

```bash

mkdir -p tasks/new_task

touch tasks/new_task/__init__.py

touch tasks/new_task/main.py

```

2. **Implement main logic** (`tasks/new_task/main.py`):

```python

from tasks.config import config

from tasks.sync_utils import create_syncer

class NewTaskRunner:

def run(self):

# Task logic

content = self.generate_content()

# Sync to platforms

syncer = create_syncer()

syncer.sync_content(

title="new-task-report",

content=content,

targets=["obsidian", "feishu"],

)

def main():

runner = NewTaskRunner()

runner.run()

if __name__ == "__main__":

main()

```

3. **Create launchd plist** (`tasks/launchd/com.zhimeng.new-task.plist`):

```xml

<?xml version="1.0" encoding="UTF-8"?>

<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "...">

<plist version="1.0">

<dict>

<key>Label</key>

<string>com.zhimeng.new-task</string>

<key>ProgramArguments</key>

<array>

<string>/path/to/virtualenv/bin/python</string>

<string>-m</string>

<string>tasks.new_task.main</string>

</array>

<key>StartCalendarInterval</key>

<dict>

<key>Hour</key>

<integer>8</integer>

<key>Minute</key>

<integer>0</integer>

</dict>

<key>WorkingDirectory</key>

<string>/path/to/zhimeng-agent</string>

<key>StandardOutPath</key>

<string>/path/to/zhimeng-agent/logs/new-task.log</string>

<key>StandardErrorPath</key>

<string>/path/to/zhimeng-agent/logs/new-task.error.log</string>

</dict>

</plist>

```

4. **Update install script** to include new task in `install.sh`

Dependencies

Notes