The Agent Skills Directory

Command Execution: The skill utilizes several standard shell utilities and project-specific scripts, such as ffmpeg, node, grep, and bash. These tools are used for video processing, extracting API signatures from project files, and managing background rendering tasks within the workspace environment.
Multi-Agent Orchestration: It delegates specialized tasks like concept creation, building, and critiquing to background agents using the start_task mechanism. This parallel workflow is managed through structured briefs that define roles and constraints for each helper.
Access and Delivery Control: A role-based hierarchy is enforced where only the lead agent has the capability to deliver final files via upload_files. Helpers are explicitly instructed to avoid external communication or uploads, centralizing control over project data.
Automated Technical Gates: The workflow incorporates several validation scripts, including video-gate.mjs, brand-lint.mjs, and artifact-contract.mjs. These are used to verify that the film acts meet technical specifications and brand guidelines before proceeding to the next phase.
Process Management: The skill manages long-running rendering operations using flock for process locking and detached shell execution. This approach ensures resource stability and prevents execution timeouts during intensive media processing.

launch-film