feat: add automated community moderation

src/routes/terms/+page.svelte

@@ -11,7 +11,7 @@
 		<p class="eyebrow"><span></span> SHD NETWORK POLICY</p>
 		<h1>TERMS & <em>PRIVACY.</em></h1>
 		<p>
-			Effective July 25, 2026. These terms set the rules for using SHD.build and explain what
+			Effective July 27, 2026. These terms set the rules for using SHD.build and explain what
 			information we process to operate and secure the service.
 		</p>
 	</header>
@@ -76,6 +76,12 @@
 						or presents a safety, privacy, legal, or abuse risk. Removing content does not
 						remove related moderation, audit, or legal records when SHD.build must retain them.
 					</p>
+					<p>
+						SHD.build may screen user content with local rules and OpenAI’s moderation API
+						before or after publication. Automated classification can publish clean content or
+						hold flagged content for Community Oversight; an authorized human makes the final
+						decision on held or reported content.
+					</p>
 				</div>
 			</section>
 			<section id="conduct">
@@ -140,6 +146,31 @@
 						notification category, and email delivery defaults to off. OAuth and email
 						providers process information under their own policies.
 					</p>
+					<p>
+						When automated content screening is enabled, SHD.build sends the submitted text
+						or custom avatar image to OpenAI’s API for safety classification. SHD.build does
+						not include your password, authentication tokens, or private sign-in credentials
+						in that moderation request. We retain the resulting category scores, flags,
+						request reference, content hash, local-rule matches, and review outcome as needed
+						to operate moderation, investigate abuse, and document decisions.
+					</p>
+					<p>
+						After SHD.build sends content to OpenAI, OpenAI controls its processing under
+						its then-current service terms, privacy policy, data controls, and SHD.build’s
+						OpenAI account settings. Those rules may permit retention, human safety review,
+						or use for model improvement, including when data sharing has been enabled.
+						OpenAI currently states that API data is not used to train or improve its models
+						unless the API customer opts in, and that default abuse-monitoring logs may be
+						retained for up to 30 days, subject to stated exceptions. SHD.build cannot
+						guarantee that OpenAI’s terms or practices will remain unchanged. Review
+						OpenAI’s
+						<a href="https://developers.openai.com/api/docs/guides/your-data#default-usage-policies-by-endpoint" rel="external"
+							>API data controls</a
+						>,
+						<a href="https://openai.com/policies/services-agreement/" rel="external">Services Agreement</a>,
+						and <a href="https://openai.com/policies/privacy-policy/" rel="external">privacy policy</a>
+						for the current rules governing API content.
+					</p>
 				</div>
 			</section>
 			<section id="cookies">
@@ -231,10 +262,11 @@
 				<div>
 					<h2>THIRD-PARTY SERVICES</h2>
 					<p>
-						SHD.build may connect to Discord, email delivery services, and future identity
-						providers. Each provider applies its own terms. Links to other sites do not imply
-						endorsement. When you choose a linked sign-in method, you authorize SHD.build to
-						request the minimum provider data needed for that method.
+						SHD.build may connect to OpenAI for content safety classification, Discord and
+						future identity providers for sign-in, and email services for message delivery.
+						Each provider applies its own terms. Links to other sites do not imply endorsement.
+						When you choose a linked sign-in method, you authorize SHD.build to request the
+						minimum provider data needed for that method.
 					</p>
 				</div>
 			</section>