The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the user to use the --break-system-packages flag with pip3 during setup. This bypasses PEP 668 protections on modern Linux distributions, which are intended to prevent conflicts between the system package manager and Python's pip. While common in containerized environments, it can affect host system stability.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of the modal library and various Python dependencies from a local requirements.txt file.
[DYNAMIC_EXECUTION]: The skill uses modal deploy to send local application code from the docker/ directory to the Modal cloud platform. This effectively creates and runs remote serverless functions controlled by the skill's code.
[INDIRECT_PROMPT_INJECTION]: The toolkit processes a text brief provided by the user to generate voiceovers, images, and video clips. This input is interpolated into command-line prompts for cloud-based AI models.
Ingestion points: User input in VOICEOVER-SCRIPT.md and configuration values in demo-config.ts.
Boundary markers: None present in the instruction examples.
Capability inventory: Scripts perform subprocess calls to AI generation tools which interact with remote APIs.
Sanitization: No sanitization or escaping is mentioned for the user-provided text before it is passed to shell commands.

video-toolkit