The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch and analyze external content from brand websites and user-provided reference media. This untrusted data is then used to generate marketing briefs, scripts, and creative assets.
Ingestion points: Website fetching for brand research and the arcads_analyze_media tool for reference ad analysis.
Boundary markers: The instructions lack explicit requirements for using boundary markers or delimitation (e.g., XML tags or Markdown blocks) when interpolating the fetched website content into prompts.
Capability inventory: The skill possesses significant generative capabilities, including video and image generation, audio/speech synthesis via arcads_audio_driven, and video editing.
Sanitization: There are no specific instructions for the agent to sanitize, escape, or filter the external content before using it to generate spoken scripts or visual overlays, creating a risk where malicious instructions embedded in a website could influence the final creative output.

ads