chengfeng-videocut
chengfeng-videocut-skills
An agent plugin for editing talking-head video: removing verbal slips, subtitles, on-screen graphics and rendering the final cut
High risk
We rate an entry high when the tool writes to external systems, handles money, production databases or secrets, or runs arbitrary commands. The CLI installs it only with your consent.
Why this level
- Downloads and runs a separate Runtime and installs background user services
- Processes and writes local media files, and bug reporting creates a GitHub issue after confirmation
Install
Manual install
node bin/install.cjs install --dry-run
node bin/install.cjs install
node bin/install.cjs doctorRun inside the plugin source directory. Requires Node.js 18+, Git and a Codex CLI that supports the codex plugin command. After installing, prepare the compatible Runtime separately.
This is third-party code. Review the repository files before installing.
What it does
The plugin adds six task-facing skills for talking-head video editing: chengfeng-cut removes slips and filler and builds a reviewed edit list without generating media, chengfeng-subtitle derives subtitle timings from that list, chengfeng-visual prepares on-screen graphics as HTML modules and styles, and chengfeng-export bakes the cut, zoom-ins, subtitles and graphics into a single mp4. Separate skills handle safe bug reporting after confirmation and update checks only through the official marketplace mechanism. The actual editing runs on a separate Runtime that the plugin downloads and prepares on instruction, while the plugin gives the agent the methodology and confirmation tools. A bundled MCP server starts together with the plugin.
Who it is for. For creators, editors and content authors who record talking-head videos and want to hand the rough cut to an agent.
Good fit when
- You need to cut slips and pauses from a talking-head recording
- You need subtitles and on-screen graphics for an assembled edit
- You need to render the project into a finished mp4 from the agent
Not a fit when
- You cannot install and run the separate Runtime that performs the editing
- You need complex cinematic post-production rather than talking-head editing
Example request
Cut the slips from this recording, build an edit list for review, then add subtitles and render to mp4Limitations
The plugin is only an install entry and methodology; it does not edit by itself: it needs a separate Runtime pinned to the plugin version, plus Node.js, Git and a compatible agent. Automatic Runtime preparation without Bun is verified only for macOS arm64; other platforms, background services and full editing are validated separately. Documentation and the interface are in Chinese. Bug reporting calls the GitHub CLI only after the user confirms.
How to disable. Remove the chengfeng-videocut plugin through your agent's plugin mechanism and delete the installed Runtime directory if you no longer need it.
Security check
- Downloads and runs a separate Runtime and installs background user services
- Processes and writes local media files, and bug reporting creates a GitHub issue after confirmation
README in short
The README explains that the repository is the public install entry for the chengfeng-videocut plugin, while the source and releases of the Runtime and Studio live in a separate repository. The plugin pins its own version, the Runtime version and an install snapshot so installation follows an exact commit rather than moving branches. Installation runs through node bin/install.cjs with Node.js, Git and a plugin-capable agent, and the installer refuses to overwrite an existing install. Update checks split candidate discovery from the write transaction, with user confirmation and provenance verification. The plugin ships six skills and an MCP server; script and server dependencies target Node.js 20 or newer. Apache-2.0 licensed.
FAQ
Does the plugin edit video itself?
No. The plugin gives the agent the methodology and confirmation tools, while a separate Runtime does the editing and must be installed and kept at a compatible version.
What language is the documentation in?
The README, interface and skill descriptions are in Chinese; keywords and some technical fields are in English.
Related
Skills and CLI from HeyGen: the agent writes HTML compositions and renders them into MP4 videos, captions, motion graphics and decks
Blender MCP
MCP for Blender
An MCP server and Blender addon: the agent creates and edits objects, materials and scenes, pulls assets and runs Python in Blender
Official Remotion skills for making videos in React: compositions, animation, captions, maps and rendering
Agentic video studio for Claude Code, Cursor and Codex: script, voiceover, images, stock footage, music, subtitles and render