remix-v2-data-flow-review

$npx mdskill add existential-birds/beagle/remix-v2-data-flow-review

Reviews Remix v2 loaders and actions for common data flow issues.

  • Catches mutations in loaders, missing validation, and leaked server fields.
  • Depends on @remix-run/node, @remix-run/react, and route module exports.
  • Checks for wrong return helpers, v1 useTransition holdovers, and revalidation traps.
  • Outputs a structured review with issue types and references to canonical patterns.

SKILL.md

.github/skills/remix-v2-data-flow-reviewView on GitHub ↗
---
name: remix-v2-data-flow-review
description: Reviews Remix v2 loaders and actions for mutations-in-loader, missing validation, leaked server fields, wrong return helpers, v1 useTransition holdovers, and revalidation traps. Use when reviewing loader/action code in a Remix v2 codebase.
---

# Remix v2 Data Flow Code Review

Targets TypeScript route modules importing from `@remix-run/*`. See [remix-v2-data-flow](../remix-v2-data-flow/SKILL.md) for canonical patterns.

## Scope

- **In scope**: route modules under `app/routes/` exporting `loader`, `action`, `shouldRevalidate`, or `headers`; components that consume `useLoaderData`, `useActionData`, `useNavigation`, `useFetcher`, `useRevalidator`, `<Await>`.
- **Out of scope**: form ergonomics (`<Form>` markup, accessibility, `useFetcher` UI patterns) → covered by `remix-v2-forms-review`. Route module conventions, file naming, nested routing, error boundary placement → covered by `remix-v2-routing-review`.
- **Imports expected**: `@remix-run/node` (or `@remix-run/cloudflare` / `@remix-run/deno`) for server utilities; `@remix-run/react` for hooks and components.

## Quick Reference

| Issue Type | Reference |
|------------|-----------|
| Mutations in loader, missing validation, leaked server fields, throwing primitives, missing param checks | [references/loaders.md](references/loaders.md) |
| Unvalidated FormData, `json` instead of `redirect` on success, missing error case, leaked actionData | [references/actions.md](references/actions.md) |
| `useTransition` v1 holdover, missing pending state, blanket `shouldRevalidate: false`, misused `useRevalidator` | [references/revalidation.md](references/revalidation.md) |
| `defer` for already-fast data, missing `<Suspense>`, no `errorElement` on `<Await>`, awaiting what should stream | [references/defer-await.md](references/defer-await.md) |

## Review Checklist

- [ ] Data needed for first render is in `loader`, not `useEffect`
- [ ] Loaders only read; writes live in `action`
- [ ] `request.formData()` results are validated (zod/valibot/invariant) before use
- [ ] Loader/action return values are projected DTOs — no password hashes, tokens, or `internal_*` fields
- [ ] `useLoaderData<typeof loader>()` uses the type annotation form (not `as Foo`)
- [ ] 404 / auth short-circuits `throw` a `Response` (or `json`/`redirect`), never a plain `Error` or string
- [ ] Successful action returns `redirect(...)` (PRG); validation failures return `json({ errors }, { status: 400 })`
- [ ] Action handles both success and error branches; no silent `return null`
- [ ] `params.foo` is checked with `invariant` / zod before use
- [ ] Pending UI reads `useNavigation()` / `fetcher.state` — no `useTransition`
- [ ] `formMethod` comparisons use UPPERCASE (`"POST"`, not `"post"`)
- [ ] `shouldRevalidate` returns `defaultShouldRevalidate` by default; opt-outs are narrow and justified
- [ ] `defer()` is used only when at least one promise streams (no `await` before passing it)
- [ ] Every `<Await>` is wrapped in `<Suspense>` and has an `errorElement`
- [ ] `useRevalidator().revalidate()` is reserved for focus/polling/SSE — not called immediately after a `<Form>` post or `fetcher.submit` (Remix already revalidates).

## Valid Patterns (Do NOT Flag)

These are correct Remix v2 usage and must not be reported as issues:

- **`useEffect` for client-only data** — Loaders run server-side; `localStorage`, `window` dimensions, `IntersectionObserver`, and browser-only APIs belong in `useEffect`.
- **`loader` returning `null`** — A loader may legitimately return `null` (e.g. optional resource not present); flag only if it should be a 404 `throw`.
- **`useLoaderData<typeof loader>()` as type annotation** — The `<typeof loader>` is a generic parameter feeding `SerializeFrom<T>`, not a `as`-style type assertion. Do not flag it as "unsafe cast."
- **Bare `new Response(body, init)` returns** — v2 routes may return any `Response`; `json()` is an ergonomic wrapper, not a requirement. Non-JSON bodies (binary, text, streams) correctly skip `json()`.
- **`return redirect(...)` from an action** — Both `return redirect(...)` and `throw redirect(...)` are legal in actions; throwing is required only from non-action helpers when you want to exit the calling function.
- **`loader` declared without the `request` arg** — Loaders may destructure only what they need (`{ params }`, `{ context }`, or `()` with no args); the unused arg is not a bug.
- **Parent `loader` revalidated after an unrelated action** — This is default Remix behavior, not a smell. Flag only if `shouldRevalidate` exists and is wrong.
- **Action returning `json({ errors }, { status: 400 })`** — This is the canonical validation-error pattern (keeps the form route rendered with field errors). Not the same as the "no redirect on success" anti-pattern.
- **`useRevalidator` for focus / polling / cross-tab sync** — These are the documented use cases; only flag manual `revalidate()` calls that immediately follow a `<Form>` post or `fetcher.submit` Remix would already revalidate.
- **`SerializeFrom`-induced type changes** — `Date` typed as `string`, `Map` typed as `{}` after deserialization is correct wire-format behavior, not a typing bug.

## Context-Sensitive Rules

Only flag these issues when the specific context applies:

| Issue | Flag ONLY IF |
|-------|--------------|
| Missing loader (using `useEffect` instead) | Data is available server-side and is NOT a browser-only API read |
| `loader` returns a raw ORM object | The object contains fields a reviewer would not paste into a screenshot (passwords, tokens, internal flags) |
| Action returns `json` on success | The action is invoked via `<Form>` causing a URL change — NOT via `useFetcher` |
| Missing pending UI | No `nav.state` / `fetcher.state` reference exists elsewhere in the file driving the same surface |
| `shouldRevalidate` returns `false` | The body has no condition or never references `formAction` / `currentParams` / `nextParams` |
| Manual `useRevalidator().revalidate()` | The call follows a Remix-managed mutation (`<Form>` post, `fetcher.submit`) — not focus / polling / websocket |
| `defer()` used | Every promise in the `defer({...})` payload was already `await`ed before the call |

## Hard gates (before writing findings)

Run these in order. **Do not draft user-facing findings until every gate passes** for the batch you are about to report.

1. **Location evidence** — **Pass:** Each issue lists the repo path to the route module and either a line range or a short verbatim quote from the file you read (not from memory or diff-only guesswork). Loader/action issues without a path to the `export async function loader|action` are not reportable.

2. **Exemption check** — **Pass:** For each issue, you can state in one line why it is *not* covered by [Valid Patterns (Do NOT Flag)](#valid-patterns-do-not-flag). In particular: confirm `useEffect` is not loading client-only data; confirm a bare `Response` return is not intentionally non-JSON; confirm a `loader` returning `null` is not a legitimate optional read.

3. **Type-annotation vs type-assertion check** — **Pass:** Before flagging an "unsafe cast" on loader/action consumption, confirm the code uses `as` (assertion) — not `useLoaderData<typeof loader>()` (annotation) and not `useActionData<typeof action>()` (annotation). The generic form is the documented safe path and must not be flagged.

4. **v1 holdover check** — **Pass:** Before flagging "missing pending state," grep the file for `useTransition`, `transition.submission`, `fetcher.type`, `formMethod === "post"` or `formMethod==='post'` (lowercase, any whitespace/quote variation), and `LoaderArgs` / `ActionArgs`. If present, the finding is a v1-holdover migration issue, not a missing-feature issue — label it accordingly.

5. **Protocol** — **Pass:** You completed the Pre-Report Verification Checklist in [review-verification-protocol](../../../beagle-core/skills/review-verification-protocol/SKILL.md) for this review.

## When to Load References

- Reviewing a `loader` body, return shape, params, throws, or sensitive-field leaks → [references/loaders.md](references/loaders.md)
- Reviewing an `action` body, FormData validation, success/error branches, or PRG redirect → [references/actions.md](references/actions.md)
- Reviewing `useNavigation` / `useTransition` migrations, `shouldRevalidate`, or `useRevalidator` use → [references/revalidation.md](references/revalidation.md)
- Reviewing `defer()`, `<Await>`, `<Suspense>`, or streaming decisions → [references/defer-await.md](references/defer-await.md)

## Review Questions

1. Is data needed for first render fetched in a `loader`, or is it stuck in a `useEffect` that defeats SSR and revalidation?
2. Does every loader return a projected DTO, or do raw ORM records (with `password`, `token`, `internal_*` fields) leak to the browser?
3. Does every action validate `request.formData()` with a schema before touching the database?
4. Does the success branch of each action `redirect(...)` so refresh / back behaves correctly (PRG)?
5. Is the consumer code using `useLoaderData<typeof loader>()` (annotation) — not `useLoaderData() as Foo` (assertion)?
6. Do any v1 holdovers remain (`useTransition`, `transition.submission`, `fetcher.type`, lowercase `formMethod`, `LoaderArgs` / `ActionArgs`)?
7. Does `shouldRevalidate` return a literal `false`, or does it reach for `defaultShouldRevalidate` and opt out narrowly?
8. Is `defer()` used only when at least one promise is passed unresolved, and is every `<Await>` wrapped in `<Suspense>` with an `errorElement`?

## Additional Documentation

- Canonical Remix v2 data-flow patterns and v1 → v2 diff → [remix-v2-data-flow](../remix-v2-data-flow/SKILL.md)
- Pre-report verification checklist → [review-verification-protocol](../../../beagle-core/skills/review-verification-protocol/SKILL.md)

## Before Submitting Findings

Complete [Hard gates](#hard-gates-before-writing-findings) (especially gate 5), then report only issues that still pass the [review-verification-protocol](../../../beagle-core/skills/review-verification-protocol/SKILL.md) pre-report checks.

More from existential-birds/beagle

SkillDescription
adr-decision-extractionUse when you need to mine a conversation, session transcript, or design discussion for architectural decisions before writing ADRs. Identifies problem-solution pairs, trade-off debates, technology choices, and explicit \"[ADR]\" tags. Triggers on \"what decisions did we make\", \"extract decisions from this chat\", \"find the choices in our discussion\", or \"summarize architectural decisions\". Also useful after long planning sessions to capture decisions that were made implicitly. Does NOT write ADR documents \u2014 use adr-writing or write-adr for that.
adr-writingUse when writing or formatting an ADR document using the MADR template, applying Definition of Done (E.C.A.D.R.) criteria, or verifying ADR completeness. Triggers on \"write the ADR\", \"format as MADR\", \"check ADR quality\", \"mark gaps in ADR\". Also triggers when a decision has been extracted and needs to become a document. Does NOT extract decisions from conversations (use adr-decision-extraction) or orchestrate the full extract-confirm-write workflow (use write-adr).
agent-architecture-analysisUse when auditing an agent codebase against the 12-Factor Agents methodology, reviewing LLM-powered system architecture, or assessing agentic app compliance. Triggers on \"analyze agent architecture\", \"12-factor audit\", \"how compliant is this agent\", or \"evaluate this LLM app\". Also applies when comparing frameworks or planning agent improvements. Not for quick checklists \u2014 this performs deep per-factor codebase analysis with file-level evidence.
app-intents-code-reviewReviews App Intents code for intent structure, entities, shortcuts, and parameters. Use when reviewing code with import AppIntents, @AppIntent, AppEntity, AppShortcutsProvider, or @Parameter.
artifact-analysisUse when the user wants a cited, structured read of local documents and project knowledge. Triggers on: \"analyze these docs\", \"scan my project for context\", \"read the docs folder\", \"summarize what's in .beagle/concepts/\", \"extract context from docs/\", \"what's in this folder\", \"go read everything in X and tell me what's there\". Also invoked programmatically by other beagle skills (prfaq-beagle Ignition, brainstorm-beagle reference points, strategy-interview context grounding) via the companion contract. Does NOT trigger on codebase lookups (\"find this function\", \"search the repo\"), web research (use web-research), LLM-as-judge evaluation (use llm-judge), or document editing (use humanize-beagle). Produces a written scan plan, parallel-subagent findings, and a cited synthesis report on disk — never inline prose, never unsourced claims.
axum-code-reviewReviews axum web framework code for routing patterns, extractor usage, middleware, state management, and error handling. Use when reviewing Rust code that uses axum, tower, or hyper for HTTP services. Covers axum 0.7+ patterns including State, Path, Query, Json extractors.
brainstorm-beagleUse when the user has a fuzzy idea and wants to shape it into a concrete project spec before planning or building. Triggers on: \"brainstorm this\", \"I have an idea for...\", \"help me think through this project\", \"what should I build\", \"spec this out\". Also catches vague feature descriptions needing structured questioning to clarify scope. Does NOT write code, plan implementation, review strategy docs, or run strategy interviews \u2014 produces a WHAT/WHY spec through dialogue, not a HOW plan.
bubbletea-code-reviewReviews BubbleTea TUI code for proper Elm architecture, model/update/view patterns, and Lipgloss styling. Use when reviewing terminal UI code using charmbracelet/bubbletea.
cloudkit-code-reviewReviews CloudKit code for container setup, record handling, subscriptions, and sharing patterns. Use when reviewing code with import CloudKit, CKContainer, CKRecord, CKShare, or CKSubscription.
combine-code-reviewReviews Combine framework code for memory leaks, operator misuse, and error handling. Use when reviewing code with import Combine, AnyPublisher, @Published, PassthroughSubject, or CurrentValueSubject.