QCut Toolkit

Unified QCut media toolkit — organize project files, process media with FFmpeg, generate AI content, control the QCut editor with native CLI commands, genera...

This package looks like a full QCut toolset (FFmpeg docs, AI pipeline, local HTTP APIs and helper scripts). Before installing or enabling it: 1) Confirm you run it inside the QCut/Electron environment it targets — otherwise expected binaries (aicp, bundled AICP binary, ffmpeg) and the local Claude HTTP server may not exist. 2) Review the bundled scripts (subtitle_server.js, review_server.js, shell scripts) before running — they can open local endpoints or access files. 3) Be aware the docs instruct persistent storage of API keys (e.g., ~/.config/video-ai-studio/credentials.env) and call external model endpoints (fal.run, Google providers). Only provide API keys you trust and consider scope-limited keys. 4) If you do not want the skill to access local files or start servers, do not run the CLI/server components; if you need to run them, sandbox or inspect them first. 5) If you want a definitive benign/malicious determination, provide the contents of the included JS/shell scripts for review (the manifest lists them but their code was not fully shown here).

Type: OpenClaw Skill Name: qcut-toolkit Version: 1.0.1 The toolkit contains several high-risk behaviors and architectural patterns that, while functional, create significant security and privacy concerns. Most notably, 'subtitles/SKILL.md' and 'talk-edit/SKILL.md' instruct the agent to upload user audio files to 'uguu.se' (a public temporary file hosting service) to facilitate transcription via the Volcengine API. The toolkit also includes local Node.js HTTP servers ('subtitle_server.js' and 'review_server.js') that execute shell commands using 'execSync' and 'spawn', presenting a potential Remote Code Execution (RCE) surface if the local ports are exposed. Furthermore, the 'videocut/self-evolve/SKILL.md' instructions explicitly direct the AI agent to modify its own skill files and rules based on 'feedback,' which could be leveraged as a mechanism for persistence or to inject malicious instructions into the agent's logic over time.