The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use shell commands including ffmpeg for audio extraction, base64 for data encoding, and curl for interacting with the OpenRouter API. These are standard utilities used to support the skill's primary media processing workflow.
[INDIRECT_PROMPT_INJECTION]: The skill processes transcripts derived from external audio and video files, which presents a surface for instructions embedded in the lecture material to influence the agent.
Ingestion points: Media transcripts saved as local text files (e.g., Lecture 05.transcript.txt).
Boundary markers: The workflow does not include specific delimiters or instructions to ignore potential commands within the transcript during the note-generation phase.
Capability inventory: The skill uses curl, ffmpeg, and jq via shell execution to handle input data.
Sanitization: No validation or sanitization of the transcript text is specified before it is processed for note creation.

lecture-notes