The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input through several tool parameters, creating a surface for potential injection attacks. Ingestion points: User-provided strings are passed to the prompt parameter in generative tools (e.g., genai-text2image, genai-text2text-completions) and datasets are ingested via the data and data_file parameters in the Variable Data API tools. Boundary markers: The instructions do not specify the use of delimiters or specific warnings to prevent the agent from following instructions embedded within the user-provided media prompts or data files. Capability inventory: The skill facilitates network operations to the Picsart API infrastructure and performs complex media transformations. Sanitization: The documentation does not define client-side sanitization of the input strings before they are transmitted to the API.
[DATA_EXFILTRATION]: The skill transmits media and prompt data to external APIs hosted by the vendor Picsart. Network requests are directed to legitimate vendor resources: mcp.picsart.io, api.picsart.io, video-api.picsart.io, genai-api.picsart.io, and vd-api.picsart.io. This activity is central to the skill's functionality.

picsart-api