refactor: earn flow + remove xstate/store by petar-omni · Pull Request #551 · stakekit/widget

petar-omni · 2026-06-26T23:13:26Z

Note

High Risk
Large new DeFi borrow path with wallet signing, transaction submission, and LTV validation touches money-moving flows; earn routing/state refactor could regress staking if mis-wired.

Overview
Adds a feature-flagged borrow experience in the dashboard (form → review → steps → complete, plus borrow position management), backed by a new borrow module: Effect Schema domain models, OpenAPI-generated BorrowApi client, and @effect/atom-react atoms for markets/positions, form state, action execution (sign/submit/confirm), and post-tx cache refresh.

Earn flow cleanup: drops EarnPageStateUsageBoundaryProvider / @xstate/store, removes common/get-token-balances.ts and getInitialToken from stake types, and tightens validator handling via a composite Validator.key (address + optional subnet) used in select-validator and position balance keys.

Tooling/config: OpenAPI generator gains BorrowApi (full httpclient + spec prep), optional CLI spec selection, VITE_BORROW_API_URL / VITE_FORCE_BORROW, and root pnpm.patchedDependencies removed; adds skeleton line/circle loaders for borrow UI loading states.

^{Reviewed by Cursor Bugbot for commit 6cf3187. Configure here.}

changeset-bot · 2026-06-26T23:13:30Z

⚠️ No Changeset found

Latest commit: ddd1bab

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

coderabbitai · 2026-06-26T23:13:34Z

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: f2d2a83c-e736-49e8-b48f-3041b4042c82

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

Create PR with unit tests
Commit unit tests in branch feat/earn-effect-atom-poc

_{Comment @coderabbitai help to get the list of available commands.}

aws-amplify-eu-central-1 · 2026-06-26T23:15:52Z

This pull request is automatically being deployed by Amplify Hosting (learn more).

Access this pull request here: https://pr-551.d2ribjy8evqo6h.amplifyapp.com

aws-amplify-eu-central-1 · 2026-06-26T23:16:01Z

This pull request is automatically being deployed by Amplify Hosting (learn more).

Access this pull request here: https://pr-551.df4xyoi0xyeak.amplifyapp.com

Key validators by address and subnet so selection keeps distinct validator options. Scope balance and init token options to the selected dashboard category.

cursor

Stale comment

cursor

Stale comment

Risk: high. Not approving: this large earn/borrow refactor exceeds the low-risk approval threshold, Cursor Security Agent left an unresolved medium-severity finding, and required checks did not finish on the latest commit (Bugbot cancelled; Security Agent still running). Assigned jdomingos and dnehl for human review.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

cursor

Stale comment

Risk: high. Not approving: this large earn/borrow refactor exceeds the low-risk approval threshold, and Cursor Security Agent has an unresolved medium-severity tabnabbing finding on complete.tsx. jdomingos and dnehl are already assigned for human review.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

dnehl

Played around with the flow - looks good.

no console errors
no "strange" multiple backend calls
duration looks good

Split earn intent and view atoms so route and wallet state changes update the resolved view. Keep dispatch separate from view reads and add wallet resolution handling. Cover dashboard tab selection and patch Effect parent tracking.

Use the native TypeScript 7 compiler across the widget and examples. Keep TypeScript 6 API compatibility for Next.js tooling.

cursor

Stale comment

Risk: high. Not approving: this large earn/borrow refactor exceeds the low-risk approval threshold, and Cursor Security Agent has an unresolved medium-severity tabnabbing finding on complete.tsx. jdomingos and dnehl are already assigned for human review.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

Replace legacy context, query, and state-machine integrations with Effect services and atom-backed wallet, API, and workflow state. Split Vitest projects and add AST rules to validate the new architecture.

cursor

Stale comment

Risk: high. Left a non-blocking comment: this earn/borrow refactor exceeds the low-risk approval threshold, and Cursor Security Agent still has unresolved high and medium findings that need human review. No new reviewers assigned; two reviewers are already requested.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

Identify GEO_LOCATION responses through Schema instead of a type predicate so extra payload fields stay valid. Drop the unused rich-error change stream and inline HTTP client setup in transport.

cursor

Stale comment

Risk: high. Left a non-blocking comment: this earn/borrow refactor exceeds the low-risk approval threshold, and Cursor Security Agent still has unresolved high and medium findings that need human review. No new reviewers assigned; two reviewers are already requested.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

Move token contracts into a host-facing Schema so Widget Configuration can strip unknown tokens and omit invalid values without rejecting an update. Recover cosmos connector load failures and keep @cosmos-kit in one package chunk so hosts do not see uninitialized live bindings. BREAKING CHANGE: SKTheme is a closed partial of the published token contract instead of an open string map.

HelpModal beside SKApp called useTranslation before WidgetTranslation existed, so titles stuck as raw keys. Render host chrome as SKApp children under the same registry, gate only the modal, and drop TrackingContextProvider. BREAKING CHANGE: TrackingContextProvider is removed. Place HelpModal as SKApp children. Bundled renderSKWidget takes SKHostConfiguration, not React children. Co-authored-by: Cursor <cursoragent@cursor.com>

cursor

Stale comment

Risk: high. Left a non-blocking comment: this earn/borrow refactor exceeds the low-risk approval threshold, and Cursor Security Agent still has unresolved high and medium findings that need human review. No new reviewers assigned; two reviewers are already requested.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

Source tests import src and miss crashes that only appear after a host rebundles dist. Add pnpm test:smoke to mount the built package and bundle in Vite and Next production hosts, and run it as a sibling CI job. Co-authored-by: Cursor <cursoragent@cursor.com>

cursor

Stale comment

Risk: high. Left a non-blocking comment: Cursor Security Agent did not reach a terminal state within the polling window, and this earn/borrow refactor exceeds the low-risk approval threshold. Human review is needed; reviewers were already assigned.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

CodeQL flagged console.log of apiKey.source because resolveApiKey returned the secret and its origin on the same object. Log the source label separately and return only the key string. Co-authored-by: Cursor <cursoragent@cursor.com>

cursor

Stale comment

Risk: high. Left a non-blocking comment: Cursor Security Agent did not reach a terminal state within the polling window, and this large earn-flow refactor exceeds the low-risk approval threshold. Reviewers were already assigned, so none were added.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

Replace React-owned history fetch hooks with focused reward-rate and TVL chart views so range switches keep the drawn series and refresh chrome can delay without changing the published load facts.

Replace generated and historical documentation with a small living set. Keep code and tests authoritative for behavior and retain only current decisions. Remove completed scratch specifications after human review.

Vite already inlines VITE_* values from the Amplify process environment. Echoing them into packages/widget/.env is redundant and turns unset variables into empty strings that skip URL fallbacks.

cursor

Stale comment

Risk: high. Left a non-blocking comment: Cursor Security Agent did not reach a terminal state within the polling window, and this earn/borrow refactor exceeds the low-risk approval threshold. Reviewers were already assigned, so none were added.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

Derive canonical networks from the generated Legacy and Yield schema intersection. Move wallet-specific definitions into the wallet domain. Use Yield project networks as the enabled-wallet source. Keep empty projects bootstrappable without exposing fallback connectors.

Enable @vitejs/plugin-react's compiler option and drop the Babel Rolldown preset. This removes @rolldown/plugin-babel and babel-plugin-react-compiler from the widget toolchain.

Recognize Robinhood mainnet and testnet as enabled EVM wallet networks. Expose their public chain IDs and viem metadata for wallet connections.

cursor

Stale comment

Risk: high. Left a non-blocking comment: Cursor Security Agent did not reach a terminal state within the polling window, and this earn-flow refactor exceeds the low-risk approval threshold. Reviewers were already assigned, so none were added.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

@effect

Track the current npm rc tag for effect and the @effect packages we already consume. Exclude those packages from the three-day release-age gate so rc.112 can be installed the day it is published.

Keep the Effect source on-disk as searchable reference for agents. Editor excludes still hide .repos from human search and auto-import.

git-subtree-dir: .repos/effect git-subtree-split: 2600f62f4532026928454dcea8d1c48557b3f942

…fect'

Replace React context and effect-based height/width sync with element slot Atoms and ResizeObserver measurements. Keep last height across route remounts so the shell does not jump, and drop the root-element portal used only for number-input sizing. Enable Biome's React Compiler lint and unblock compiler-incompatible render patterns in wallet scope, virtual list, and Ledger Live navigation.

cursor

Stale comment

Risk: high. Left a non-blocking comment: this earn-flow refactor exceeds the low-risk approval threshold, and Cursor Security Agent still has unresolved high findings that need human review. Reviewers were already assigned, so none were added.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

Stay on the snake_case 1.x line so cosmos-kit chain types remain compatible.

cursor

Stale comment

Risk: high. Left a non-blocking comment: Cursor Security Agent did not reach a terminal state within the polling window, and this earn-flow refactor exceeds the low-risk approval threshold. Reviewers were already assigned, so none were added.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

The API `reason` is English prose with the actual amounts. Use it as the details line when the Widget Instance is English, and keep catalog copy for French so untranslated backend text never leaks.

cursor

Risk: high. Left a non-blocking comment and did not approve: this large earn-flow refactor exceeds the low-risk threshold, and Cursor Security Agent did not finish within the wait window. Human review is needed; no additional reviewers were assigned because two are already requested.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

petar-omni added 2 commits June 27, 2026 01:10

feat(widget): add effect atom earn flow

8f6710c

refactor(widget): replace xstate stores with effect atoms

15d72af

fix(widget): stabilize earn option resolution

3127f0d

Key validators by address and subnet so selection keeps distinct validator options. Scope balance and init token options to the selected dashboard category.

petar-omni marked this pull request as ready for review July 7, 2026 12:22

Philippoes previously approved these changes Jul 7, 2026

View reviewed changes

cursor Bot reviewed Jul 7, 2026

View reviewed changes

Comment thread packages/widget/src/features/borrow-transaction-flow/ui/complete.tsx

feat: borrow init

5dac944

petar-omni dismissed Philippoes’s stale review via 5dac944 July 7, 2026 12:27

petar-omni force-pushed the feat/earn-effect-atom-poc branch from 6cf3187 to 5dac944 Compare July 7, 2026 12:27

petar-omni requested a review from Philippoes July 7, 2026 12:27

Philippoes previously approved these changes Jul 7, 2026

View reviewed changes

feat: temp

2110846

petar-omni dismissed Philippoes’s stale review via 2110846 July 7, 2026 12:31

cursor Bot requested a review from dnehl July 7, 2026 12:31

cursor Bot reviewed Jul 7, 2026

View reviewed changes

cursor Bot requested a review from jdomingos July 7, 2026 12:31

cursor Bot reviewed Jul 7, 2026

View reviewed changes

dnehl previously approved these changes Jul 7, 2026

View reviewed changes

dnehl self-requested a review July 7, 2026 14:11

petar-omni added 3 commits July 9, 2026 16:56

fix(widget): stabilize earn atom subscriptions

90a2b21

Split earn intent and view atoms so route and wallet state changes update the resolved view. Keep dispatch separate from view reads and add wallet resolution handling. Cover dashboard tab selection and patch Effect parent tracking.

feat: schema decoding, query to atoms migration

b2043f0

build(typescript): migrate to typescript 7

3c5bfdc

Use the native TypeScript 7 compiler across the widget and examples. Keep TypeScript 6 API compatibility for Next.js tooling.

petar-omni dismissed dnehl’s stale review via 16d2bb4 July 14, 2026 23:49

cursor Bot reviewed Jul 14, 2026

View reviewed changes

refactor(widget): migrate runtime state to effect atoms

67eb76f

Replace legacy context, query, and state-machine integrations with Effect services and atom-backed wallet, API, and workflow state. Split Vitest projects and add AST rules to validate the new architecture.

petar-omni force-pushed the feat/earn-effect-atom-poc branch from 16d2bb4 to 67eb76f Compare July 15, 2026 08:42