The Agent Skills Directory

[SAFE]: The skill provides a highly structured production pipeline using 'hard gates' (approval checkpoints). This design ensures that the AI agent waits for explicit user confirmation at every stage (Intake, Concept, Scenes, Shot Design, Production, Post) before proceeding, which is a robust safety and usability practice.
[SAFE]: All external URLs included in the reference chapters are provided as legitimate citations for filmmaking theory, AI model documentation, and industry news. No code-loading or automated script execution from these domains is attempted.
[SAFE]: The skill utilizes a role-play persona ('Award-winning director') that includes instructions to prioritize storytelling over immediate visual generation. These instructions (e.g., 'stop the user from generating pixels before the story is locked') are creative constraints intended to improve output quality and do not attempt to bypass platform safety filters.
[PROMPT_INJECTION]: The skill includes instructions to 'challenge, don't please' the user and to 'refuse to skip ahead' if the user requests immediate generation. These are task-specific behavioral constraints designed to enforce the filmmaking workflow and are not malicious injections.
[EXTERNAL_DOWNLOADS]: Numerous links to well-known AI services (ElevenLabs, Midjourney, Google DeepMind, Runway) and established industry resources (StudioBinder, No Film School, Wikipedia) are present. These are documented as reference material and citations, posing no risk to the execution environment.

director