The Agent Skills Directory

[SAFE]: The skill provides templates for using curl to interact with Resemble AI's official API endpoints at app.resemble.ai. All network activity is directed to the vendor's infrastructure for the stated purpose of media analysis.
[SAFE]: Security instructions are explicitly included, mandating the use of environment variables for the RESEMBLE_API_KEY and warning the agent never to print bearer tokens or secrets in chat outputs.
[COMMAND_EXECUTION]: The skill provides shell command templates for the agent to perform media uploads and poll for results. This includes a benign Python snippet used for parsing JSON responses in environments where jq is unavailable.
[SAFE]: A "Red Flags" section provides behavioral constraints for the agent, such as preventing the agent from guessing media authenticity and ensuring it only reports completed results from the API.
[SAFE]: The skill handles potentially sensitive media by promoting the use of zero_retention_mode, which instructs the API to delete submitted content after analysis.

resemble-detect