All jobs

Add a 404 Page to a Vue Router App

Yes. An agent adds a catch-all route and a NotFound view to a Vue 3 app, sends a missing article there without changing the URL, tests it and opens a PR.

Run this on RushOpens the console with this prompt ready to send, in a Project that has your repository.

What the agent does

The agent reads the router before it adds a route. In an app shaped like realworld-apps/vue-realworld-example-app, package.json pins vue ^3.5, vue-router ^5.1, pinia ^3 and vite ^8, with @playwright/test ^1.60, eslint and prettier for development, installed with bun; the scripts are serve, build, lint, format:check and test:e2e. src/router/index.js declares ten named routes, home, home-tag, login, register, settings, profile, profile-favorites, article, editor-new and article-edit, each as a lazy import() of a view without the .vue extension, which vite.config.js allows through resolve.extensions, and meta.requiresAuth on the settings and editor routes. There is no catch-all, so an unknown URL renders App.vue's header and footer around an empty router-view. src/main.js registers a router.beforeEach that awaits auth.checkAuth() and redirects to /login when a guarded route is hit without a session. src/views/Article.vue fetches the article and its comments in beforeRouteEnter through the Pinia article store and swallows a failure with .catch(() => null), so a missing slug renders the page with an empty article. src/views/Login.vue shows the view style: defineOptions({ name: 'RwvLogin' }), Conduit markup from the realworld submodule's theme, script setup.

It adds src/views/NotFound.vue in that style: RwvNotFound, a container page with a heading, the path that was requested, and a router-link to home. src/router/index.js gets a last entry named not-found on /:pathMatch(.*)*, lazily importing the view like the others and without requiresAuth, so the guard in main.js lets it through once checkAuth() has resolved. Article.vue stops swallowing a 404: beforeRouteEnter keeps tolerating a comments failure, but when fetchArticle rejects with error.response.status of 404 it returns the not-found route with pathMatch set from to.path, so the address bar keeps the article URL and the user sees the 404 view instead of an empty banner. Other failures still render as before. The header stays as it is, and nothing in the auth store changes.

playwright.config.ts points testDir at ./realworld/specs/e2e, the shared suite in the submodule, and starts bun run serve on 127.0.0.1:8080. The agent does not edit the submodule; it adds e2e/not-found.spec.ts in the repository and a second projects entry in playwright.config.ts with testDir: './e2e', so the shared suite runs unchanged beside the new file. The spec visits a made-up path and expects the heading and the home link, visits /article/no-such-article and expects the same heading with the URL unchanged, and visits a guarded path while signed out to confirm the redirect to /login still wins. It runs bun run lint and bun run format:check, the two steps in .github/workflows/lint.yml, bun run build, and bunx playwright install --with-deps chromium followed by bun run test:e2e as .github/workflows/playwright.yml does, and pastes the output in the pull request.

Start it with one API call

POST https://api.phoenix.so/o/{org}/sessions with this body:

{
  "agent": "claude",
  "input": "Add a 404 page to this Vue 3 app and open a pull request. Read package.json, vite.config.js, src/main.js, src/App.vue, src/router/index.js, src/views/Login.vue, src/views/Article.vue, src/store/article.js, src/common/api.service.js, playwright.config.ts and the two workflows under .github/workflows first. Add src/views/NotFound.vue named RwvNotFound in the style of Login.vue with a heading, the requested path and a router-link to home. Add a last route named not-found on /:pathMatch(.*)* that lazily imports it, with no requiresAuth. In Article.vue's beforeRouteEnter, when fetchArticle rejects with a 404, return the not-found route with pathMatch built from to.path so the URL stays; keep other failures as they are. Add e2e/not-found.spec.ts and a second Playwright project with testDir ./e2e, leaving the realworld submodule suite untouched, covering an unknown path, a missing article with the URL unchanged, and a guarded path still redirecting to /login. Run bun run lint, bun run format:check, bun run build and bun run test:e2e and paste the output in the pull request. Do not change the header, the auth store or the submodule.",
  "resources": [
    {
      "url": "https://github.com/your-org/your-app"
    }
  ]
}
  1. Get an API key. In the console, open Settings, then API keys, and create a key for your org. Send it in the Authorization header as a Bearer token.
  2. Send the request. POST the body in request.json to https://api.phoenix.so/o/{org}/sessions, with your repository as the resources URL. The run starts in its own cloud sandbox.
  3. Read what came back. Follow the run on its live stream or wait for the run.completed webhook, then read the result and its trajectory.

What comes back

What comes back is a pull request that adds src/views/NotFound.vue and e2e/not-found.spec.ts, edits src/router/index.js, src/views/Article.vue and playwright.config.ts, and leaves main.js, the header and the stores alone. The description shows the three Playwright cases and pastes the eslint, prettier, vite build and Playwright output.

On a repository with the Format Check and Playwright Tests workflows, the pull request runs the lint steps and the end-to-end suite with the submodule checked out, so the new project runs beside the shared specs.

A reviewer checks that the catch-all is the last route, that the article redirect only fires on a 404 and not on a network failure, that the requested path is rendered as text, and that a signed-out visit to a guarded URL still lands on the login page. Then they decide whether the server or CDN should answer unknown paths with a real 404 status.

Run it on every Vue app with a bare router

A team with several Vue front ends sends one POST /o/{org}/sessions per repository with this prompt and that repository in resources. Each run reads its own router, so an app on createWebHashHistory gets the same catch-all and an app with nested layouts gets the view inside the right parent, and each run opens its own pull request.

Give the API key a webhook_url and Rush POSTs run.completed when each turn ends, with the session_url to read the pull request and the test output from.

The use case this job follows, with its automation and trigger: Hand off a long task.

Why run it on Rush?

A 404 page is only right in the browser: the catch-all must lose to every real route, the guard must still redirect, and a missing article must land on the page without rewriting the address. In a Rush sandbox the agent starts the Vite server, drives the three cases through Playwright in Chromium and keeps the output in the run's record, so the pull request shows the behaviour rather than describing it.

The change is one view, one route entry and one branch in beforeRouteEnter, in the shapes the app already uses, so there is no dependency and nothing to configure. The same request works on any Vue 3 app on vue-router.

Frequently asked questions

Does an unknown URL return a real 404 status?

Not from the app; a single-page app answers every path with the same index.html and the router decides what to render. The view is the 404 the user sees. A real status needs the host to map unknown paths, and the pull request notes that as a deployment step.

Why keep the URL when an article is missing?

So the address bar still shows what was asked for and a reload or a bookmark behaves the same. beforeRouteEnter returns the not-found route with pathMatch set from the original path, which renders the view without a redirect in the history.

Does the catch-all interfere with the auth guard?

No. The guard in main.js runs on every navigation and only redirects routes with requiresAuth or the following feed. The not-found route has no such flag, so it renders for signed-in and signed-out visitors alike.

Why a second Playwright project instead of a new spec in the suite?

The suite lives in the realworld submodule, which the repository tracks by commit and does not edit. A second project with its own testDir runs the repository's own spec beside the shared ones under the same webServer.

More Vue jobs

The same job on other stacks

Keep reading