The Agent Skills Directory

[SAFE]: All network communication is restricted to the official Pachca API at api.pachca.com. This is consistent with the skill's purpose and is performed using standard curl commands.
[SAFE]: The skill does not contain hardcoded credentials. It explicitly instructs the agent to request an access token from the user if one is not already provided in the context.
[SAFE]: The file upload process implements a secure two-step mechanism. It first retrieves upload parameters and a pre-signed direct_url from the Pachca API before performing the actual file transfer, which is a standard security pattern for cloud storage.
[SAFE]: Although the skill processes untrusted data when reading message history, this is an inherent feature of a messaging integration. The risk is minimized as the agent's capabilities are focused on API interactions within the Pachca ecosystem.

pachca-messages