Media Processor

音视频处理器 - 企业级多媒体内容处理工具 | Media Processor - Enterprise multimedia content processing

Security Scan

The package and runtime instructions are internally consistent with an on‑host multimedia processing tool; it asks for no unrelated credentials or privileged persistence and contains only local media-processing code (FFmpeg/Whisper).

Purpose & Capability

Name/description (media processing) matches the included code (video/audio editor, ffmpeg wrapper, whisper-based transcriber) and requirements (FFmpeg, moviepy, whisper). No unrelated credentials, binaries, or config paths are requested.

Instruction Scope

SKILL.md instructs only local media processing and how to install dependencies. The transcriber loads Whisper models at runtime — that may cause large model downloads from the network (expected for transcription). The runtime instructions do not reference unrelated files, environment variables, or external endpoints beyond the normal model download behavior.

Install Mechanism

There is no automated install spec; SKILL.md instructs using pip install -r requirements.txt and installing FFmpeg manually. This is reasonable for a Python media tool, but note that installing whisper/torch can trigger large downloads and GPU drivers; model weights may be downloaded at first use.

Credentials

The skill requires no environment variables or credentials. Declared dependencies and external requirements (FFmpeg, Whisper) are proportionate to transcription and transcoding functionality.

Persistence & Privilege

The skill does not request always:true, does not modify other skills, and contains no code that writes to system-wide agent configs. It runs as a normal user tool operating on local files.

Assessment

This code appears to be a straightforward on-host media processing library. Before installing or running: 1) ensure FFmpeg and ffprobe are installed and on PATH (video_processor uses ffprobe); 2) be aware Whisper (openai-whisper) may download large model weights at first use and require significant CPU/GPU and disk space; 3) the code has a few implementation bugs (e.g., video_processor sometimes appends '-c:v copy' as a single arg which can break ffmpeg invocation, and tests expect an 'error' key that get_info currently doesn't return) — these are quality issues but not indicators of malicious behavior; 4) run code on non-sensitive/test media first and inspect any network activity (model downloads) if you need to enforce offline operation; 5) if you will process untrusted filenames, validate/normalize paths (they are written into a temporary concat list file and may be interpreted by ffmpeg).

Media Processor - 音视频处理器

企业级多媒体内容处理解决方案，支持音视频转码、剪辑、转录和格式转换。

核心功能

功能模块	说明
格式转换	支持 50+ 种音视频格式互转
视频剪辑	裁剪、合并、添加水印、调整分辨率
音频处理	降噪、音量调整、格式转换、片段提取
智能转录	语音转文字（支持中英文）
批量处理	多文件并行处理，支持队列

快速开始

from scripts.video_processor import VideoProcessor

# 视频转码
processor = VideoProcessor()
processor.convert('input.mp4', 'output.webm', 
                 video_codec='vp9', audio_codec='opus')

# 视频剪辑
processor.clip('input.mp4', 'output.mp4', start='00:01:30', duration=60)

安装

pip install -r requirements.txt
# 确保系统已安装 FFmpeg
ffmpeg -version

项目结构

media-processor/
├── SKILL.md                 # Skill说明文档
├── README.md                # 完整文档
├── requirements.txt         # 依赖列表
├── scripts/                 # 核心模块
│   ├── video_processor.py   # 视频处理器
│   ├── audio_processor.py   # 音频处理器
│   ├── transcribe_engine.py # 转录引擎
│   └── format_converter.py  # 格式转换器
├── examples/                # 使用示例
│   └── basic_usage.py
└── tests/                   # 单元测试
    └── test_processor.py

运行测试

cd tests
python test_processor.py