The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides implementation patterns that utilize system-level command execution to handle media processing.
Evidence: rules/audio-sync.md and rules/web-export.md contain Python and JavaScript snippets that use subprocess.run and execSync to call ffmpeg for video muxing and encoding.
[DYNAMIC_EXECUTION]: The skill encourages the use of interactive debugging features which open an embedded shell within the execution environment.
Evidence: rules/community-edition-patterns.md and rules/opengl-renderer-patterns.md document the use of interactive_embed() and embed(), allowing for real-time code execution and object inspection.
[PROMPT_INJECTION]: The skill contains an attack surface for indirect prompt injection as it processes untrusted text data for video and audio generation.
Ingestion points: rules/audio-sync.md defines workflows where user-supplied script text is passed to Text-to-Speech (TTS) engines and Manim text objects.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands were found in the templates.
Capability inventory: The skill has access to subprocess execution (ffmpeg), file system writes/deletes, and interactive Python shells.
Sanitization: No sanitization or validation of input script content is implemented.
[EXTERNAL_DOWNLOADS]: The skill references several external libraries and well-known service providers for media enhancement.
Evidence: rules/audio-sync.md lists integrations with well-known TTS services including OpenAI, Azure, and ElevenLabs. It also references the edge-tts and pydub Python packages.

motion-graphics