The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several Bash, Python, and Node.js scripts to automate the video production workflow. These include scripts/batch-generate.sh for coordinating tasks via a CLI, scripts/analyze-beats.py for audio processing using standard libraries (librosa), and scripts/generate-preview.mjs for creating HTML storyboards. All command executions are functional and scoped to the skill's intended purpose.
[EXTERNAL_DOWNLOADS]: The skill uses curl in scripts/batch-generate.sh and references/stage-generate.md to download generated video assets from the platform. These downloads are directed toward the service's official infrastructure and represent legitimate operational requirements.
[DATA_EXFILTRATION]: No evidence of unauthorized data access or exfiltration. The skill operates on project-specific files and user-provided creative materials, with network activity limited to the platform's API and asset retrieval.
[PROMPT_INJECTION]: The instructions are focused on guiding the AI's creative behavior and pipeline management. There are no patterns suggesting attempts to bypass safety filters, extract system prompts, or override core agent instructions.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data such as user briefs and source material (Mode D). While it lacks explicit boundary markers for these inputs, the subsequent processing (generating video prompts for a specific CLI) presents a low risk of exploitation within the context of video generation.

director