YouTube Intelligence Extractor
PassAudited by ClawScan on May 15, 2026.
Overview
This appears to be a straightforward YouTube transcript extraction skill, with notable but purpose-aligned use of external transcript fetching, an optional Python package fallback, and automatic Markdown report creation.
This skill looks safe for normal YouTube transcript extraction. Before using it, be aware that it may query a third-party transcript mirror, may install a Python package if the web method fails, and will automatically save a Markdown report unless you instruct the agent otherwise.
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.
The agent may contact Wave Tube or search the web using the YouTube video information to retrieve a transcript.
The skill directs the agent to fetch transcript content from an external transcript mirror. This is aligned with the YouTube transcript extraction purpose, but users should know a third-party site may be queried.
Use the `web_fetch` tool on the Wave Tube URL.
Use this skill for public or intended-to-share YouTube videos; avoid using it for private or sensitive video links unless you are comfortable with third-party fetching.
If the fallback is used locally, the agent may install a package into the Python environment, which could affect other Python tools.
The fallback method includes installing an unpinned Python package and uses --break-system-packages. This is purpose-aligned for transcript fetching, but it can modify the local Python environment.
pip install youtube-transcript-api --break-system-packages -q
Prefer running this fallback in a virtual environment or approve the package install explicitly before allowing it.
Using the skill may create Markdown report files without a separate confirmation step.
The skill instructs automatic file creation and presentation after completing a report. This is disclosed and related to the output workflow, but it is still an automatic file-writing behavior.
After presenting the report, always save it as a markdown file and present it to the user. This is standing behavior — do not wait to be asked.
Check where the agent saves generated reports and ask it to avoid saving files if you only want an in-chat answer.