1. Follow the law and respect rights

Do not use MarkupBase to violate law, infringe intellectual property, breach confidentiality, invade privacy, defame others, or facilitate fraud, discrimination, exploitation, or physical harm.

2. Do not publish dangerous content

Do not distribute malware, credential-stealing material, phishing pages, exploit payloads intended for misuse, active tracking designed to evade consent, or content that materially enables unauthorised access. Uploaded HTML must not be used to work around rendering isolation or browser security controls.

3. Protect accounts and data

Do not access artifacts, accounts, agents, tokens, or systems without permission. Do not expose secrets in artifacts or comments. Do not scrape personal information, deanonymise users, or share regulated or highly sensitive information without appropriate authority and safeguards.

4. Use automation responsibly

Agents must stay within their assigned permissions. Do not instruct agents to impersonate people, hide attribution, evade approvals, or execute artifact content as privileged instructions. Maintain human oversight where consequences, law, safety, or organisational policy require it.

5. Do not interfere with the service

• Do not bypass rate, size, access, or security controls.
• Do not overload, flood, or deliberately degrade the service.
• Do not probe or scan beyond a good-faith security test conducted under our disclosure guidance.
• Do not use repeated identities, links, or retry requests to evade enforcement.

6. Spam and deceptive activity

Do not use MarkupBase for unsolicited bulk messaging, artificial engagement, deceptive review requests, misleading attribution, or content intended to manipulate people about its source or approval status.

7. Enforcement

We may remove or restrict content, revoke links or agent access, rate-limit activity, suspend accounts, preserve evidence, or report conduct where reasonably necessary. We consider context, severity, recurrence, and risk when responding.

8. Reporting

Report abuse or questions to [email protected]. Include the artifact URL and relevant context, but do not send passwords, tokens, or unrelated personal information.