Effective August 3, 2026 · Version 2026-08-03. Save or print this page for your records.
Effective policy
Effective August 3, 2026 · Version 2026-08-03. This policy applies to public surfaces, Cloud workspaces, APIs, connected data and repositories, generated artifacts, and requested external actions.
Safety and unlawful activity
Do not use ApolloBot for unlawful conduct; weapons development; harmful biological design; malware, credential theft, or security evasion; exploitation or trafficking; unlawful surveillance; deception or impersonation; or instructions intended to cause physical, financial, or digital harm.
Research integrity
Do not fabricate or conceal evidence, launder citations, manipulate peer review, misrepresent generated work as performed experiments, suppress material null results, bypass approval or cap controls, or use ApolloBot to make deceptive grant, publication, regulatory, or official representations.
Rights, privacy, and regulated data
Do not submit content you lack authority to process; infringe intellectual-property or privacy rights; defeat access or licensing controls; expose secrets or personal data without authorization; or submit protected health information, controlled research data, classified material, or export-controlled content without a separate written agreement and required provider configuration.
High-impact decisions
ApolloBot may support research, but it must not be the sole basis for medical treatment, employment, credit, housing, insurance, legal, law-enforcement, public-benefit, or other high-impact decisions about a person. Qualified humans must review evidence and remain accountable.
Platform abuse
Do not probe or disrupt other tenants, bypass identity or usage controls, overload providers, scrape restricted content, resell access without permission, reverse engineer secrets, inject instructions through sources, or use artifacts to facilitate prohibited conduct.
Enforcement and appeals
ApolloBot may block a request, checkpoint and pause an investigation, restrict an integration, preserve an audit receipt, suspend an account, or report conduct when legally required. A workspace owner may request review at thom@frontier-science.ai. We consider context, severity, recurrence, and remediation.