The Agent Skills Directory

[SAFE]: The skill follows established best practices for its stated purpose. It correctly utilizes platform-specific tools for media generation and project management without attempting to bypass security controls or exfiltrate sensitive data.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting untrusted data from brand research briefs and voice-of-customer dossiers (advibly_get_brand, advibly_get_brand_dossier) which are then interpolated into prompts for image and voiceover generation. This is considered a functional design characteristic for personalizing ad content.
Ingestion points: External brand identity and research data are fetched via advibly_get_brand and advibly_get_brand_dossier (SKILL.md).
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are used when interpolating external brand data into generation prompts.
Capability inventory: Extensive generative capabilities including advibly_generate_image, advibly_generate_video, advibly_generate_voiceover, and advibly_generate_music, alongside file management and social media posting tools.
Sanitization: No evidence of sanitization or validation of the ingested brand data before its use in downstream prompts.

claymation-ad