Use the agent
with judgment.
Chusky helps people and products move work forward. These plain-language principles describe the responsibilities that come with using an agent that can read, create, and act.
The basics
You are responsible for the work you authorize.
Use lawful inputs
Do not use Chusky to access data, systems, or accounts without permission.
Review important actions
Treat agent output as assistance. Check consequential messages, decisions, documents, and changes before they leave your control.
Protect your keys
Keep account credentials, project keys, and connected-service access out of public repositories and browser bundles.
Agent boundaries
Automation does not remove accountability.
Do not rely on Chusky as a substitute for professional advice, emergency response, safety-critical control, or a person with the authority to make a regulated decision.
Do not attempt to bypass approval boundaries, impersonate another person, interfere with a service, or use generated content to mislead people about its origin.
We may limit or suspend access when use creates a security, safety, legal, or reliability risk to other users or the service.
Changes and support
The product will keep evolving.
If a workflow matters, keep the relevant identifiers, approvals, and outputs in your own systems as well. That makes your product resilient when an external service changes.