chatgpt-video-editing-skills

Why we included this project

You can turn an agent that already works with your local files and terminal into a small video pipeline with this pair of skills, one for setting up the editing environment and one for doing the actual edit. The editor follows eight documented steps, from inspecting the source and pulling word-level transcript timestamps, through strategy approval, rough cuts, and caption burn-in, to preview, QA, and the final render. Its safety habits are the real draw: the agent asks before uploading footage or spending paid credits, never overwrites the original material, and treats a file as finished only after decoding and QA checks pass. Captions are built for Traditional Chinese using Source Han Sans, and the pipeline wires together FFmpeg and video-use, with ElevenLabs transcription or HyperFrames animation available when an approved strategy needs them. For a creator already comfortable directing an agent, this offers a repeatable route to captioned vertical shorts without opening a dedicated editor.