Devotd Github
ReleaseDeveloper tools

kajisho5/ffmpeg-skill v2.22.1 — rejecting protocol-prefixed inputs and outputs

kajisho5/ffmpeg-skill v2.22.1 now rejects input and output values carrying a protocol prefix, preventing the mismatch between how URLs and paths were interpreted.

Repositorykajisho5/ffmpeg-skill ↗

kajisho5/ffmpeg-skill v2.22.1 now rejects input and output values carrying a protocol prefix, preventing the mismatch between how URLs and paths were interpreted.

GitHub description: Local video editor for coding agents: 42 FFmpeg tools with a machine-readable contract and an MCP server, for Claude Code, Cursor and Codex. Cut, captions, reframe, loudness, platform exports. No cloud, no API keys.

stars
★ 4.1K
Forks
322
Language
Python
License
MIT
Latest release
v2.22.1
Read original source ↗View on GitHub ↗

kajisho5/ffmpeg-skill is a repository of local video editing tools for Claude Code, Cursor, and Codex. v2.22.1, published on the GitHub releases page, now rejects input and output values that carry a protocol prefix. According to the release notes, this is a change meant to prevent the problem where values read as URLs and values read as paths were interpreted inconsistently. The point of this update is not the addition of video editing features but making the tools refuse values with that prefix.

The repository provides 42 FFmpeg tools, machine-readable contracts, and an MCP server. Its features include trimming video, subtitles, screen re-layout, volume adjustment, and per-platform export. Because it is a project that bundles local video editing so that it can be used from a coding agent, the way input and output values are interpreted is directly tied to how the tools behave. This release changes one of them: whether values with a protocol prefix are accepted.

However, the provided release notes do not specify which protocol prefixes are rejected or which tools the rule applies to. They also give no examples of cases where a URL and a path were previously interpreted inconsistently, nor the scope of the change's impact. So the only thing that can be confirmed about this change is that the intent was to reject such values and prevent interpretation mismatches; how far it reaches into real usage is hard to judge from the release information alone.