The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a Python script (scripts/upload_asset.py) that uses the requests library to interact with Cloudinary's REST API. The script is designed for local execution by the agent to perform media uploads, transformations, and management tasks.
[EXTERNAL_DOWNLOADS]: The skill communicates with official Cloudinary API and CDN domains (api.cloudinary.com and res.cloudinary.com). These are well-known services required for the skill's stated purpose of media management.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests external data (user-provided media files and API responses containing metadata).
Ingestion points: Media files uploaded by the user and JSON responses from Cloudinary's Admin and Upload APIs (including AI-generated tags).
Boundary markers: None explicitly defined in the prompt templates for separating API data from instructions.
Capability inventory: The skill can read local files and perform network requests to Cloudinary.
Sanitization: The skill uses structured JSON parsing for API responses, which provides inherent structure but does not specifically sanitize text fields (like tags or metadata) that might be interpreted as instructions by the agent.
[CREDENTIALS_SAFE]: The skill follows security best practices by instructing the user to store sensitive API credentials (API Key, Secret, Cloud Name) in environment variables or a CLOUDINARY_URL string rather than hardcoding them.

cloudinary-assets