docs(readme): add deployment FAQ (#379)

* docs(readme): add deployment FAQ

Add a FAQ section right after the deploy steps, covering the four most
commonly asked deployment questions:

- GitHub flags the account when the fork reuses this project's name, so
  Cloudflare cannot list the repository (404)
- redeploy after deleting the Worker still asks for an invite code,
  because D1 / KV were not deleted
- JWT_SECRET set as a build variable instead of a runtime secret
- JWT_SECRET stored as a plain text variable, so it is lost on upgrade

* docs(readme): clarify FAQ and deployment instructions

README.md

@@ -91,6 +91,25 @@
 >   | KV | No | 25 MiB (Cloudflare limit) | 1 GB |
 
 
+## FAQ
+
+- **After forking the repository, why can't I see my repository when connecting GitHub to Cloudflare, or why do I get a 404 after selecting it?**  
+  This is usually related to how the GitHub fork is identified or how Cloudflare handles repository authorization and synchronization. If the fork keeps a repository name, description, or other information that is very similar to the upstream project, it may be more likely to trigger related restrictions or issues. It is recommended to rename the repository to something different from the upstream project when creating the fork, and change the repository description as well. For example, you can rename it to `2233warden`. If you have already created the fork, you can rename the repository and update its description in the GitHub repository settings, then try connecting it to Cloudflare again.
+
+- **I deleted my deployment and redeployed it. Why does registration require an invite code again?**  
+  Deleting the Worker or redeploying it does not automatically delete the persistent data that was already created. The users, invite codes, and related configuration stored in the D1 database and KV namespace are still there, so the newly deployed Worker continues to read the existing data and enforce the invite-code requirement.  
+  If you want to start completely from scratch, you need to delete the corresponding **D1 database and KV namespace** as well.
+
+- **I configured `JWT_SECRET`, but the page still says it is missing. Why?**  
+  Make sure `JWT_SECRET` is configured under **Workers → Settings → Variables and Secrets**, specifically as a **Runtime variable or Secret**, rather than under **Build variables**.  
+  Build-time variables are only available during the build process. They are not available to the Worker at runtime, so the build may succeed while the application still reports that `JWT_SECRET` is missing.
+
+- **Why does `JWT_SECRET` seem to disappear after an upgrade or redeployment?**  
+  It is recommended to store `JWT_SECRET` as a **Secret** rather than as a plain-text variable. `JWT_SECRET` is a sensitive runtime credential and should not be committed to the repository.  
+  If your deployment process recreates or overwrites the Worker variable configuration, ordinary variables may be affected. Secrets are more appropriate for sensitive configuration that needs to remain available across multiple deployments. If the application still reports that `JWT_SECRET` is missing after a redeployment, check **Variables and Secrets** for the current Worker and make sure the Secret is still configured.
+
+---
+
 ## How to update
 
 - Manual: open your fork on GitHub; when the sync banner appears, click **Sync fork** → **Update branch**