The Agent Skills Directory

[SAFE]: The skill implements a robust set of workflow rules for interacting with a research-focused MCP server. It emphasizes using server-provided identifiers, following recommended tool sequences, and adhering to status messages.
[SAFE]: Data handling for sensitive operations, such as payments and report sharing, is managed securely by relying on URLs and summary information provided directly by the tool responses rather than having the agent construct them.
[SAFE]: The use of base64 for image uploads is a standard technical requirement for the cookiy_media_upload tool and does not represent an attempt to obfuscate commands or malicious content.
[SAFE]: The skill incorporates defensive programming patterns such as optimistic locking via base_revision and idempotency keys to ensure reliable and safe tool execution.

cookiy-workflows