Community contract
Rules that keep the signal useful
Last updated: 15 August 2026
Keep it useful
Share concrete experiences with Codex, questions, working methods, projects and startups. Project sharing is welcome in the matching category when it explains what was built, what help is needed or what others can learn. Repetitive promotion, referral campaigns and context-free advertising are spam.
No personal attacks or targeting
Do not publish threats, hate speech, private information, coordinated harassment, impersonation, employee targeting, malware or deceptive links.
External links
Submit only a destination you are entitled to share and describe it honestly. Project and repository fields accept direct public HTTPS addresses and reject common shorteners, credentials, unusual ports, local hosts and raw IP addresses. Human review and the exit warning cannot prove trust or prevent later compromise. The External Links Policy applies to every visit.
Protect private data
Do not post passwords, access tokens, private keys, recovery codes, payment details, government identifiers, precise home addresses, private communications, exploit secrets or another person's data without lawful authority. Redact logs and screenshots before sharing. Never ask another member to move to a private channel to collect credentials or money.
Rights and accuracy
Share only material you created, have permission to use or may lawfully quote. Do not impersonate a person or project, fabricate endorsements, hide affiliate interests, or make knowingly false safety claims. Project owners are responsible for keeping their destination accurate and reporting a compromise.
Standard signal and terminal bonus
The RESET button grants one standard signal per anonymous signed browser actor and active cycle. Finding and running the terminal easter-egg command grants one separate bonus signal in that cycle. Repeating either route does not increase the counter again. Both are community requests—not commands sent to OpenAI and not proof of unique physical people.
Moderation
New comments and replies enter a review queue. When enabled, AI may check safety, relevance and likely duplicates, but it does not own the comment. A suggested rewrite is published only after the author explicitly accepts it; otherwise a human moderator may approve or reject the original. If AI is unavailable, the comment remains safely queued and rejected content is not shown publicly.
Live chat is a separate stream
Live-chat rooms, messages and rules are separate from the comment archive. Chat uses deterministic length, repetition, mention, link, slow-mode and operator-managed term rules plus human reports and sanctions. AI does not classify, hide or approve live-chat messages. The labelled Reset Agent may reply only after an explicit @ResetAgent mention in a room where the operator enabled it; it has no moderation or admin authority.
Reset Agent answers
Reset Agent is a read-only explainer, not an official OpenAI source. Its answers are limited to current D1 aggregates, Community Pulse and saved source evidence. The UI shows the server-verified basis and freshness separately; missing evidence must be stated, and agent answers never change votes, cycles, comments or moderation decisions.
Helpful votes
Each browser actor may mark a published comment as helpful once and may remove that mark. Self-voting is blocked by the server. Helpful votes are community relevance signals, not payment, rank or moderation privileges.
Reporting and enforcement
Use the Legal & Safety Report for suspicious links, unlawful content, privacy, copyright, security or AI concerns. Include the exact content or page and a factual explanation. RWReset may preserve evidence, temporarily restrict content, request more detail, reject an insufficient or abusive notice, or remove material. Knowingly false reports and retaliation are prohibited.
Consequences
Depending on severity and repetition, RWReset may warn, reject, label, hide or remove content; disable a link; slow or block submissions; restrict an actor; or preserve evidence required for a dispute. Moderation is applied in good faith but cannot guarantee that every violation is found immediately.
Independent project
Are We Reset Yet? is community-built and is not affiliated with, endorsed by or operated by OpenAI.