learn-from-video
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill offers to install necessary browser automation tools if they are not detected in the environment. These include official packages from well-known providers such as Playwright (Microsoft), Puppeteer (Model Context Protocol), Browserbase, and Browser-use. The installation is user-triggered and uses standard package managers (npx, pip).
- [COMMAND_EXECUTION]: The skill executes shell commands to perform its core functions:
- It runs a local Python script
skills/learn-from-video/scripts/extract-transcript.pyto retrieve video transcripts usingyt-dlporyoutube-transcript-api. - It executes package manager commands (
npx,pip) to install browser automation tools if authorized by the user. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data from YouTube transcripts and metadata to generate learning segments and questions. This represents a potential injection surface if the external content contains malicious instructions designed to influence the agent's behavior during the session.
- Ingestion points: Video transcripts and metadata extracted from YouTube via browser scraping or API helpers (SKILL.md Phase 2).
- Boundary markers: No explicit delimiters or instructions to ignore embedded commands within the transcript data are defined in the instructions.
- Capability inventory: The skill possesses the ability to execute shell commands, control a web browser, and write Markdown files to the local file system.
- Sanitization: No specific sanitization or filtering logic is mentioned for the ingested transcript content before it is used for question generation.
Audit Metadata