LEGAL // ACCEPTABLE USE POLICY

Acceptable Use Policy

Effective: November 14, 2026 · Keoir Enterprises, LLC

sgt.sh relays SSH traffic from you and your AI agents to servers. That is a powerful capability, and this policy draws the line on how it may be used. It applies to everyone who uses the Service and to every agent acting under an account, and it is part of the Terms of Service. You are responsible for your agents' conduct as if it were your own.

1. Only systems you are authorized to access

You may register as an Endpoint, connect to, or run commands on a system only if you own it or have the owner's explicit permission to access and administer it. The following are strictly prohibited:

  • Accessing, probing, or attempting to access any system without authorization.
  • Credential stuffing, password guessing, or using credentials you are not entitled to use.
  • Using the Service to maintain access to a system after your authorization has ended.
  • Penetration testing of third-party systems, even "for research", unless you hold written authorization from the system owner covering the activity and the time window.

2. No attacks and no abuse of the network

You may not use the Service — including its shared and dedicated egress IPs — to:

  • Launch denial-of-service attacks, port scans, or sweeps of third-party networks.
  • Send spam, phishing messages, or other unsolicited bulk communications.
  • Distribute malware, operate botnet or malicious command-and-control infrastructure, or host material that facilitates attacks on others.
  • Conceal the origin of attacks or route traffic through our relays to evade blocks, sanctions, or law enforcement.

Egress IPs are shared infrastructure with a reputation that every customer depends on. Activity that gets our address space blocklisted is treated as abuse even if the target did not complain.

3. No unlawful use

  • No use of the Service to violate any applicable law or regulation, including computer crime, export control, sanctions, and privacy laws.
  • No storing or transmitting content through the Service that you do not have the right to process.

4. Agents and automation

  • Scope agents to the narrowest set of endpoints that does the job; use command policies and approvals for destructive operations.
  • Do not configure agents to approve their own actions or to bypass the human-approval controls the Service provides.
  • You must be able to stop your agents. If an agent of yours is misbehaving, disable the agent or use the kill switch — and tell us if it affected systems beyond your own.

5. Protect the Service itself

  • No attempts to breach, scan, or stress-test sgt.sh infrastructure, to escape session isolation, or to access other tenants' data. (Found a vulnerability? We want to hear about it: me@keoir.dev.)
  • No sharing, reselling, or white-labeling access to the Service without our written agreement.
  • No circumventing plan limits, metering, or per-IP session caps.
  • No excessive automated load unrelated to legitimate use of your own endpoints.

6. Enforcement

We may investigate suspected violations and may suspend sessions, agents, endpoints, or entire accounts — immediately and without prior notice where the violation creates risk to others, to the law, or to the Service. We prefer to warn first when it is safe to do so. Where the law requires it, we report illegal activity to the authorities.

7. Reporting abuse

To report abuse originating from sgt.sh infrastructure (our egress ranges are published on the site), email me@keoir.dev with timestamps, source IPs, and logs. We act on verified reports promptly.