Magic Link Sign-In
Sign in from an emailed link instead of remembering a password.
What it adds
Passwordless authentication by email: a short-lived single-use link, purpose-bound tokens, and defined behaviour when the link opens somewhere unexpected.
What your agent is told to do
5
What your agent is told to do
5-
1
Make links short-lived — minutes, not hours — single-use, and bound to the purpose they were issued for. A sign-in token must not also work as a reset or invite token.
-
2
Bind the request to the browser that started it: issue a paired value on the request page and require it, so a link opened on another device follows the handoff path instead of silently signing in.
-
3
Give the same response to every address entered, and never reveal whether an account exists at any step.
-
4
Validate the post-login destination against an allowlist and drop anything external.
-
5
Do not make magic links the only way in for accounts that matter. Email is a single point of failure; keep at least one other method available and say so at sign-up.
Edge cases it handles
6
Edge cases it handles
6- Opening the link on a different device or browser must offer an explicit confirmation step rather than failing or silently authenticating.
- Link scanners in corporate mail systems will hit the URL before the user does — consuming the token on GET breaks sign-in entirely for those users.
- Requesting a second link must invalidate the first, and the emails must be distinguishable so people do not click the stale one.
- A consumed or expired link must show a neutral 'request a new one' screen with no account details.
- Requests must be rate-limited per address and per IP; the endpoint sends email on demand.
- The token must not survive in browser history, referrer headers, or server logs after use.
Definition of done
8
Definition of done
8- Links expire in minutes, are single-use, and are valid only for sign-in.
- A link opened in a different browser requires explicit confirmation.
- Pre-fetching by mail scanners does not consume the token.
- Post-login redirects are allowlisted and never external.
- Responses never disclose whether an address is registered.
- An alternative sign-in method remains available on every account.
- The feature matches the existing design system.
- No existing functionality is broken.
Related features
User Roles
User Roles
Not everyone should be able to do everything.
What it does
Role-based permissions, enforced on the server, with a UI that reflects them.
How it works
-
1
Define roles and their permissions in ONE place. Scattering
if user.admin?through the codebase is how authorization bugs happen. - 2 Enforce every permission server-side, at the controller or service layer.
- 3 Hide UI the user can't use — but never rely on hiding as the enforcement.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/user-roles
Workspace Deletion Workflow
Workspace Deletion Workflow
Close a whole workspace without one click erasing everyone's work.
What it does
An owner-only deletion flow with an explicit inventory, a recovery window, and cascading cleanup across internal and external systems.
How it works
- 1 Restrict deletion to owners and require recent re-authentication plus typing the workspace name. This is the most destructive action in the product and should feel like it.
- 2 Show an inventory before confirming: how many members lose access, what happens to billing, how many files and records are affected, which integrations disconnect, and which shared links die.
- 3 Schedule the deletion rather than executing it. Mark the workspace as pending deletion, block access to it, notify every member and admin, and only purge after a stated recovery window.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/workspace-deletion-workflow
Domain Verification
Domain Verification
Prove an organization controls a domain before unlocking domain-wide powers.
What it does
A DNS or file-based challenge that verifies domain ownership and gates features like enforced SSO and domain-wide join.
How it works
- 1 Generate a unique, unguessable token per organization and domain, and verify it via a DNS TXT record or a file served over HTTPS at a fixed path.
- 2 Check the challenge server-side from the app's own infrastructure. A verification the browser performs can be faked.
- 3 Re-check verified domains on a schedule and mark a domain unverified when the evidence disappears, rather than trusting a one-time pass forever.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/domain-verification
How it works
-
1
Copy the link
Grab the Markdown instruction URL for this feature.
-
2
Give it to your AI
Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.
-
3
It inspects, then implements
Your agent reads your existing app first, then adds the feature to fit it.
Works with your stack
These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.
Need it tighter than that? Customize the feature and tell it exactly what you're running.