Clipify analyzes a source video to find high-value short moments (punchlines, reactions, quotable lines), trims candidate clips, and optionally reformats them for vertical social formats (9:16) using face-aware pan or split-screen layouts. It also generates and burns word-timed captions (opus-style) for polished social delivery.
Use when the user asks to make short clips from a longer video, requests reframing to 9:16/vertical, wants subtitle styles (opus/karaoke/minimal), or asks to auto-detect funny/highlight moments for social posting. Ideal for creators who need fast, repeatable clip generation.
Best used with multimodal/code-capable agents that can run ffmpeg/Whisper and basic Python helpers (Claude Code, Copilot/Codex-style agents).
Clipify is a video clipping skill that finds funny moments in videos, reframes to vertical formats, and burns word-timed captions using Whisper + ffmpeg. SKILL.md is well-structured with clear steps and a pitfalls section. All four scripts (analyze.py, audio_align.py, build_ass.py, build_pan.py) are functional data-processing tools but crash immediately without arguments — no argparse, no usage help, no input validation. Scripts require specific CLI args and input files to operate, which is expected for pipeline tools, but the lack of any error messaging is a quality gap.
Clean skill — no security concerns. Scripts are purpose-built pipeline components (not standalone CLIs), so missing argparse is somewhat expected but still a quality gap. SKILL.md is thorough with good progressive disclosure and practical pitfalls section.