Acceptable Use Policy

Status: in force. Written by the operators of this service, not by a lawyer.

This document was drafted in-house and has not been reviewed by counsel. We are telling you that rather than leaving you to assume otherwise, because you are entitled to know how much weight to put on it. What it does have going for it is that every factual claim about how the software behaves was checked against the implementation instead of adapted from a template — where the system does not do something a policy would normally promise, this says so.

If you are relying on this for a decision that matters, get your own advice.

Last updated: 2026-08-16.

This policy is part of the Terms of Service. It applies to showusyourprompts.com as operated by Aleph Consulting LLC, a Nevada limited liability company, not to independent instances of the software.

Do not publish

• Unlawful content, or content that infringes someone else's rights. Copyright claims have their own process — see COPYRIGHT.md.
• Prompts whose evident purpose is to cause serious harm — for example prompts designed to produce malware, credential-harvesting content, or material that facilitates violence against people.
• Sexual content involving minors. Reported to the relevant authorities, always, and the account terminated.
• Harassment, threats, or targeted abuse of a person or group.
• Impersonation of a person or organisation.
• Secrets. API keys, passwords, tokens, personal data about other people. Note that a public repository is *fully public*: publishing a secret here is publishing it to everyone, and forks of it are outside our reach.
• Spam — bulk, repetitive, or purely promotional repositories.

Adversarial and "jailbreak" prompts — the line, and why it is here

This is a prompt-sharing platform, and prompts that probe, test or document the boundaries of AI systems are legitimate subject matter with real research and safety value. Publishing a prompt that circumvents a model's safeguards is not, by itself, a breach of this policy.

What breaches it is publishing such a prompt in order to accomplish something in the "Do not publish" list above — to generate malware, abuse material, or targeted harassment. The line is purpose and effect, not technique.

Why we draw it there

Because the alternative does not make the work stop. It makes it private.

People will study how these systems fail whether or not we host it. Banning the category from a public, attributed, reviewable repository does not remove that research from the world — it removes it from the light, into places with no provenance, no licence, no moderation and no one accountable for it. A finding nobody can cite is a finding nobody can fix.

So we would rather protect the right to study, learn and improve these systems in the open. That is a considered position with a real cost, stated here so that nobody has to guess at it:

• It is harder to moderate. Judging purpose means reading context — the summary, the framing, the repository around a prompt — rather than pattern-matching on content. We accept that cost deliberately.
• We will sometimes get it wrong, in both directions. Where we do, the report and appeal paths above are the remedy, and a moderation action always carries a recorded reason.

None of this weakens anything else in this policy. A prompt whose evident purpose is to cause serious harm is prohibited regardless of how it is framed, and "it's research" is not a defence for publishing abuse material, malware, or harassment.

Do not

• Circumvent access controls, rate limits, or the platform's authorization.
• Automate access in a way that degrades the Service for others. Rate limits are enforced, and the limit and your remaining allowance are returned on every API response as the X-RateLimit-Limit and X-RateLimit-Remaining headers. We publish the mechanism rather than a number here on purpose: the number is tuned operationally and differs on self-hosted deployments, so a figure written into this policy would be wrong somewhere the day it was written. The headers are always current. Exceeding the limit returns 429; sustained, deliberate evasion of it is a breach of this policy.
• Use the Service to distribute malware.
• Misrepresent the provenance or authorship of a prompt.

Reporting

Anyone can report content, including without an account, via the in-product report path. Reports are reviewed by platform staff. See COPYRIGHT.md for infringement specifically, and SECURITY.md for vulnerabilities.

What happens on breach

Proportionate to what happened, and to whether it is repeated:

• content hidden or removed;
• account suspended or deactivated;
• referral to authorities where the law requires it.

Each moderation action is recorded with the acting staff member and a reason.