DORK / LEGAL · DRAFT FOR REVIEW

Acceptable use policy

Draft — not final policy. Effective date: [EFFECTIVE DATE]. Operator: [FULL LEGAL ENTITY NAME]. Abuse and security contact: [VERIFIED ABUSE CONTACT]. The operator, enforcement procedure, and notice process require confirmation before publication as a binding policy.

Authorized research

Use Dork for lawful research that you are authorized to conduct. Obtain permission for datasets, papers, code, systems, and personal information before submitting or testing them. An academic purpose does not remove privacy, intellectual-property, institutional, or security obligations.

Do not use Dork to conduct unauthorized access, distribute malware, steal credentials, commit fraud, exploit people, facilitate violence, or infringe others’ rights. Do not use a hosted Lab to disrupt other services, send spam, mine cryptocurrency without authorization, or operate unrelated abusive workloads.

Respect the boundaries

Do not bypass permissions, isolation, source restrictions, safety controls, usage limits, or billing controls. Do not try to extract other researchers’ data or credentials. Do not share a provider account in a way its agreement prohibits, and do not use model connections you are not authorized to use.

Security testing requires explicit permission for the target and method. This policy does not grant authorization to probe Dork, its infrastructure providers, or third-party systems. Report a suspected vulnerability to [SECURITY CONTACT AND DISCLOSURE PROCESS] without exposing private research or secrets.

Research integrity

Keep generated accounts, observed experiments, formal checks, and human judgments distinguishable. Do not fabricate citations, verification evidence, experimental results, or independent review. Do not describe unchecked material as established merely because an AI researcher states it confidently.

Review code and outputs before running them outside their intended environment, submitting them for publication, or using them in consequential decisions. Dork does not supply approval from an ethics committee, dataset owner, institution, or regulator.

Provider requirements also apply

Hosted use must comply with the applicable Google Cloud terms and Acceptable Use Policy.

ChatGPT use is subject to the applicable OpenAI Terms of Use, API or business use to the applicable Services Agreement, and both to applicable OpenAI usage policies.

Where Claude is available, follow the applicable Anthropic Consumer Terms or Commercial Terms and Usage Policy. Other configured providers impose their own requirements.

Use the version applicable to your service, jurisdiction, account, and agreement. Provider policies can change. A Dork feature or model connection is not permission to violate them, and this draft does not assert that every provider or subscription path is supported.

Reports and enforcement

Send suspected misuse to [VERIFIED ABUSE CONTACT] with the information needed to locate the issue. Exclude credentials and unrelated research. [CONFIRM INVESTIGATION, ACCESS RESTRICTION, EMERGENCY ACTION, NOTICE, APPEAL, AND DATA-PRESERVATION PROCEDURES].

Technical holds can pause work even without a policy violation. Conversely, the absence of a technical block does not make prohibited use acceptable.