The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it is designed to ingest and analyze YouTube transcripts, which are untrusted external data sources.
Ingestion points: Transcripts are retrieved via the scripts/fetch_transcript.py script or can be manually pasted by the user into the chat context.
Boundary markers: Absent. The skill instructions (SKILL.md) do not include delimiters or specific system instructions to the agent to ignore any command-like text contained within the transcripts.
Capability inventory: While the skill focuses on content analysis, an attacker could potentially use a malicious transcript to hijack the agent's broader capabilities, such as file system operations or network access, depending on the host environment.
Sanitization: Absent. The skill does not perform any filtering or sanitization on the transcript text before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill relies on an external Python dependency for its core functionality.
Dependency: youtube-transcript-api is required for the scripts/fetch_transcript.py script to function.
Context: This is a well-known and standard library used for accessing YouTube's transcript features without requiring an official API key.

youtube-video-analyst