The core idea of DeepSeek Harness (DSH) is “everything is a plugin,” and plugins can quickly extend an agent’s internet access and data processing capabilities. In local advertising delivery, tasks such as organizing customer leads, campaign plans, and video analysis data often require cross-platform copy and paste, which is time-consuming and error-prone. This page introduces the dsh-douyin-oem-touliu-report plugin, which is designed to solve the problem of automated daily and weekly report generation for OEM ad delivery.

Plugin Information

  • Name: dsh-douyin-oem-touliu-report
  • Author: ruiyukirin
  • Category: Online tools
  • License: MIT

Core Features

After installation, the plugin automatically deploys the douyin-oem-touliu-report skill to the user skills root directory (~/.dsh/skills/), and DSH immediately gains this capability. Its core features include:

  1. Data collection: Collects three-source data from Douyin Laike and Ocean Engine local delivery, including the lead center, plan-level delivery, and video analysis.
  2. Report generation: Calculates key metrics and generates an A4-layout OEM delivery daily or weekly report in .docx format.
  3. Automation support: Supports optional email notifications and Windows scheduled tasks.

Installation and Enablement

Install it using the official plugin marketplace:

dsh plugin --profile web add dsh-douyin-oem-touliu-report

After installation is complete, restart dsh web, and confirm that the plugin appears in Settings → Plugin Market. The douyin-oem-touliu-report directory should also appear in the skills directory.

Configuration and Execution

1. Business Configuration

Copy the configuration template and fill in the business information:

# 复制示例配置
cp ~/.dsh/skills/douyin-oem-touliu-report/config/config.example.json ~/.dsh/skills/douyin-oem-touliu-report/config.json

Fill in the following information in config.json:
* Company name: The enterprise identifier.
* Account IDs: advid, groupid, and view_id.
* Entry URLs: The four data entry addresses.
* Paths: The data directory and report directory (absolute paths are recommended).

2. Email Notification (Optional)

If you need to send reports automatically, configure SMTP:

cp ~/.dsh/skills/douyin-oem-touliu-report/config/.env.example ~/.dsh/skills/douyin-oem-touliu-report/config/.env

Fill in the SMTP credentials for QQ/163/Outlook and the recipient email address in config/.env.

3. Environment Dependencies

Install the Python dependencies and configure the browser driver:

pip install -r requirements.txt
playwright install

For the first run, use the local Edge browser to log in to Douyin Laike and obtain the login state. Subsequent runs can reuse this state.

4. Manual Execution

Run it in the skills directory:

cd ~/.dsh/skills/douyin-oem-touliu-report
python scripts\run.py --date 2026-08-27    # 生成指定日期的日报（默认为前一日）

Automation Configuration

Add a schedule section to config.json to set the automatic execution time for daily or weekly reports:

"schedule": {
  "daily_enabled": true,
  "daily_time": "09:30",
  "weekly_enabled": true,
  "weekly_time": "10:00",
  "weekly_day": "MON"
}

After saving, restart DSH or run setup_schedule.ps1. The system will automatically register a Windows scheduled task. The task runs silently under the current Windows user, and logs are stored in the runtime/ directory.

Notes

  • Redacted release: This package is a redacted release and does not include any real company names, account IDs, SMTP credentials, or runtime data. It must be configured manually after installation.
  • Permissions and paths: The scheduled task runs under the current Windows user. The data_dir and report_dir fields in config.json must use absolute paths; otherwise, the task may fail because the directories cannot be found.
  • Globally available: This plugin is an open-source release and can be installed with one click by users worldwide.

Summary

This plugin uses automated scripts to solve the tedious data organization work in Douyin local delivery OEM operations. For developers or operations staff who need to regularly output structured reports, it provides an end-to-end toolchain from data collection to report generation. For more details and source code, see the GitHub repository.