The Agent Skills Directory

[SAFE]: The skill implements a legitimate media production pipeline using platform-provided tools such as advibly_generate_image, advibly_generate_video, and advibly_render_composition. All operations are consistent with the skill's stated purpose of generating advertising content.
[SAFE]: The design incorporates mandatory human-in-the-loop approval stages for the cast sheet, beat map, and storyboard stills. These checkpoints provide oversight, preventing the unintended generation of content or excessive consumption of platform credits.
[PROMPT_INJECTION]: The skill features a surface for indirect prompt injection by using user-supplied brand and product descriptions to populate generative prompts. Ingestion points: Phase 1 (intake) in SKILL.md. Boundary markers: Absent. Capability inventory: Media generation and composition tools. Sanitization: Absent. The risk associated with this surface is mitigated by the mandatory manual approval steps required before final production.
[COMMAND_EXECUTION]: The references/audio-and-gotchas.md file contains examples of ffprobe commands for file validation. These are informational examples for users and do not represent a security threat.

pixar-style-ad