The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/roadshow_composer.py utilizes subprocess.run to call ffmpeg for video encoding, audio mixing, and subtitle integration. The implementation uses list-based arguments to avoid shell injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (user documents and brand materials) to extract core information and guide subsequent generation roles. 1. Ingestion points: Document analysis (Role 1) and brand style learning (scripts/style_learner.py). 2. Boundary markers: Not explicitly implemented in the skill instructions or scripts to separate untrusted content from system instructions. 3. Capability inventory: File system write access across all scripts, ffmpeg command execution in scripts/roadshow_composer.py, and network API calls for TTS in scripts/audio_processor.py. 4. Sanitization: No specialized filtering or escaping is applied to the text content extracted from user documents before it is used for content planning or script writing.
[EXTERNAL_DOWNLOADS]: The skill configuration defines dependencies on standard media processing libraries (moviepy, pillow, pydub) and the system-level FFmpeg utility.

ppt-roadshow-generator