The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were detected. The skill is intended to guide an AI agent through a structured scriptwriting process for short-form video content.
[DATA_EXFILTRATION]: The skill references reading context files from the .agents/ directory (e.g., .agents/product-marketing-context.md). These are project-specific files used for agent personalization and do not constitute an exposure of sensitive system credentials or private keys. No network exfiltration patterns were found.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from user arguments and external context files. However, the skill lacks tool-based execution capabilities (such as filesystem write or shell execution), and the agent's role is limited to text generation, mitigating the risk of exploitation.

04-script-video