1Scope
This policy applies to every agent Block9 LLC configures or manages for you, and to your use of the client portal. It forms part of our Terms of Service.
2Prohibited uses
You may not use an agent we manage to:
- Break the law, or help anyone else do so.
- Deceive people about who or what they are dealing with, including denying the agent is an AI when asked or when disclosure is legally required.
- Send unsolicited bulk messages, or contact people who have opted out, in violation of TCPA, CAN-SPAM, or similar rules.
- Record calls where you have not met your jurisdiction's consent requirements.
- Harass, threaten, defraud, discriminate against, or unfairly target any person or group.
- Give medical, legal, financial, or other professional advice outside a scope your licensed professionals have approved.
- Handle protected health information, payment card data, or other regulated categories unless we have expressly agreed in writing to the controls that requires.
- Make automated decisions with legal or similarly significant effects on a person — credit, employment, housing, insurance — without meaningful human review.
- Attempt to extract our system instructions, another client's data, credentials, or model configuration.
- Deliberately consume resources to inflate usage or degrade the service for others.
- Resell, sublicense, or expose the agent as your own product without a written agreement.
3High-risk actions
Certain actions stay behind a human approval gate by default: moving money, issuing refunds or credits, deleting or irreversibly changing records, sending mass outbound communication, and granting or changing system access.
You may ask us to automate some of these. We will do so only with your written acceptance of the residual risk, and we may decline where the risk is unreasonable — including to you.
4What we do to keep agents in bounds
- Ground every answer in knowledge you approved before launch.
- Force escalation to a human when the agent is unsure or out of scope.
- Refuse professional advice and never invent prices, availability, or policies.
- Apply least-privilege access to every connected system.
- Screen inbound messages for prompt injection, credential probing, and abuse before they reach a model.
- Log actions so there is a record of what happened and when.
These are controls, not guarantees. An agent can still be wrong, and our Terms are explicit that we do not promise otherwise.
5Enforcement
If we become aware of use that violates this policy, we will normally contact you first and work it out. Where the risk is immediate — legal exposure, harm to individuals, a threat to platform access that other clients depend on — we may suspend an agent first and explain immediately after.
Because our service depends on third-party model providers, conduct that risks our access to those platforms puts every client at risk, and we will act quickly to contain it.
6Reporting
To report misuse of an agent, contact support@block9.app or call 636-224-8069.