inSaiAI Intelligent Editing
v1.0.0Use when performing video/audio processing tasks including transcoding, filtering, streaming, metadata manipulation, or complex filtergraph operations with FFmpeg.
⭐ 0· 1.1k·2 current·2 all-time
byDonsonAI@liudu2326526
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description match the content: the SKILL.md is a comprehensive FFmpeg guide (transcoding, filtering, metadata, filter_complex examples). Nothing in the file asks for unrelated resources or credentials.
Instruction Scope
Instructions are limited to FFmpeg/FFprobe command examples and explanations, referencing expected local files (input.mp4, watermark.png, filelist.txt, etc.). There are no steps that read unrelated system files, call external endpoints, or request secrets. Note: the skill's output-format is 'shell-command', so the agent will produce shell commands an executor might run — those commands should be reviewed before execution.
Install Mechanism
No install spec and no code files — lowest-risk form. The skill assumes ffmpeg/ffprobe are available on the host but does not attempt to fetch or install anything.
Credentials
The skill declares no required environment variables, credentials, or config paths. The content does not access or require secrets beyond typical local file inputs.
Persistence & Privilege
always is false and the skill is user-invocable; it does not request persistent or elevated platform privileges and does not modify other skills or system settings.
Assessment
This skill is a text-only FFmpeg how-to and appears coherent. Before installing or running commands produced by the agent: ensure ffmpeg/ffprobe are installed in a safe environment; always review the generated shell commands before executing them (they can overwrite files or run long CPU/GPU tasks); avoid running commands on sensitive systems or with privileged accounts; when processing untrusted media, run in an isolated sandbox or VM to limit potential risks. If you want the agent to execute commands automatically, require an explicit approval step to review each command first.Like a lobster shell, security has layers — review code before you run it.
latestvk973s6f1yrjtbk96sbqv552vwd80vttv
License
MIT-0
Free to use, modify, and redistribute. No attribution required.