Video Analyzer
PassAudited by ClawScan on May 1, 2026.
Overview
The skill’s files match its stated purpose of downloading and analyzing videos, with only disclosed local processing and optional user-directed AI sharing to be aware of.
This appears safe to install for its stated purpose. Use a dedicated output folder, expect local downloads/transcription to use disk and compute resources, and be careful before piping transcripts from meetings, due-diligence videos, or other sensitive sources into an AI service.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Running the skill can download large files, use significant compute time, and replace generated files in the selected output folder.
The script downloads a user-supplied video URL, processes it with local media/transcription tools, and writes or overwrites files in the chosen output directory. This is central to the skill’s purpose, but it is real local and network tool authority.
URL="$1"; OUTPUT_DIR="${2:-./outputs}" ... yt-dlp ... "$URL" ... ffmpeg ... "$OUTPUT_DIR/audio.mp3" -y ... whisper "$AUDIO_FILE" ... --output_dir "$OUTPUT_DIR"Run it on intended video URLs only, use a dedicated output directory, and avoid pointing it at folders where similarly named files should not be overwritten.
If the transcript is pasted or piped to an AI agent, malicious or irrelevant instructions inside the transcript could affect the summary task.
The summarization helper embeds raw transcript text into an AI prompt. Transcripts from web videos or meetings are untrusted content and could contain instructions aimed at influencing the model.
Transcript: $CONTENT
When summarizing, tell the AI to treat the transcript strictly as data and not to follow any instructions contained inside it.
Sensitive meeting, podcast, or due-diligence transcript content may be sent to an AI provider depending on how the CLI is configured.
The documentation suggests an optional flow that sends transcript contents into another AI/agent CLI. This is disclosed and user-directed, but it crosses a data boundary.
cat outputs/transcript.txt | clawdbot ask "Summarize this"
Only use the AI-piping option for content you are allowed to share, and check the AI provider’s privacy and retention settings for sensitive transcripts.