feat: add is user elibile for trial method

[tuna1207]

Explanation

Clients currently have no single place to ask whether a user can start a free trial for a subscription product. Each product has its own rule:

• Shield: the user has never trialed Shield.
• Money Account Plus: the user has never trialed it and signed in with social login.

This PR adds SubscriptionController:isUserEligibleForTrial(productType), which returns a Promise<boolean>:

• If trialedProducts already includes productType, it returns false straight away.
• For Shield, it returns true without calling any other controller.
• For Money Account Plus, it returns the result of SeedlessOnboardingController:getIsUserAuthenticated. That action returns false for SRP users and for locked wallets. Only unlocked users can subscribe, so this works as the social-login check.
• The method uses a switch on productType. Adding a new member to PRODUCT_TYPES fails compilation until the switch handles it.
  To call the seedless action, @metamask/subscription-controller now depends on @metamask/seedless-onboarding-controller. It's a type-only import, and the package is already a dependency of profile-sync-controller, so this adds no circular dependency.
  @metamask/wallet now delegates SeedlessOnboardingController:getIsUserAuthenticated to the SubscriptionController messenger, so the default wallet setup keeps working.

References

Checklist

• I've updated the test suite for new or updated code as appropriate
• I've updated documentation (JSDoc, Markdown, etc.) for new or updated code as appropriate
• I've communicated my changes to consumers by updating changelogs for packages I've changed
• I've introduced breaking changes in this PR and have prepared draft pull requests for clients and consumer packages to resolve them

---

Note

Medium Risk
Breaking messenger wiring is required for all non-wallet hosts, and trial gating for Money Account Plus depends on seedless auth state being correct.

Overview
Adds SubscriptionController:isUserEligibleForTrial, a single messenger-backed way for clients to ask whether a user can start a free trial for a product.

Eligibility first checks trialedProducts in controller state; if the product was already trialed, the method returns false. Shield is eligible when that check passes. Money Account Plus also requires SeedlessOnboardingController:getIsUserAuthenticated (used as the social-login gate). Unknown product types throw at runtime, and new PRODUCT_TYPES values must be handled in the switch or TypeScript fails to compile.

This is a breaking integration change: SubscriptionControllerMessenger must be allowed to call SeedlessOnboardingController:getIsUserAuthenticated. @metamask/subscription-controller adds a dependency on @metamask/seedless-onboarding-controller, and @metamask/wallet delegates that action in its default SubscriptionController setup so trial checks work out of the box.

^{Reviewed by Cursor Bugbot for commit 9670d93. Bugbot is set up for automated code reviews on this repo. Configure here.}

[mcmire]

Just a few suggestions to the changelogs.

[mcmire]

Why is this a breaking change? Adding a new method to a class should not be breaking. It's also not common to provide implementation details in changelogs so perhaps we can simplify this a bit:

Suggested change

	- **BREAKING:** Add `SubscriptionController:isUserEligibleForTrial` to report whether a user can start a trial for a product ([#10622](https://github.com/MetaMask/core/pull/10622))
	- Shield is eligible when the user has not trialed Shield before.
	- Money Account Plus is eligible when the user has not trialed it before and `SeedlessOnboardingController:getIsUserAuthenticated` is true.
	- Add `SubscriptionController:isUserEligibleForTrial` to report whether a user can start a trial for a product ([#10622](https://github.com/MetaMask/core/pull/10622))

[tuna1207]

oh this is breaking because it requires adding new messenger action for this new method

[mcmire]

Are you saying that in order to call this method, SeedlessOnboardingController:getIsUserAuthenticated must be delegated? If so, we could combine this and the entry below. I've made a separate suggestion for that.

[tuna1207]

Yes, agree 👍

[mcmire]

Everything looks good now! Thanks for your patience.