Skip to content
Back to skills

Platform Frontend

ASecurity

"Develop and validate the AutoGPT Platform Next.js frontend,

  • 247 stars
  • 0 votes
  • 0 copies
  • 4 views
  • Added September 8, 2026
developmentpythonbashreactnextjsnodedockertestingapifrontendbackend

Works with

  • cli
  • api

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned September 8, 2026

npx -y skills add VectorSpaceLab/AREX-Skill --skill platform-frontend --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Platform Frontend?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Platform Frontend
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/vectorspacelab-platform-frontend/badge)](https://www.skillsdirectory.com/skills/vectorspacelab-platform-frontend)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: platform-frontend
description: "Develop and validate the AutoGPT Platform Next.js frontend,
  generated API hooks, Builder/Copilot UI, design system, and frontend tests."
metadata:
  disco-role: operating
disable-model-invocation: true
license: NOASSERTION
---

# Platform Frontend

Use this sub-skill for the `autogpt_platform/frontend` Next.js App Router
application: authenticated platform pages, Builder, Copilot, Library,
Marketplace, settings, teams, generated API hooks, feature flags, design-system
components, and frontend test suites.

## Reference map

- Read [frontend architecture](references/frontend-architecture.md) for route
  groups, component/hook structure, state, API clients, auth, and design rules.
- Read [API and testing](references/api-and-testing.md) for Orval generation,
  MSW handlers, Vitest integration tests, Playwright happy paths, Storybook,
  and command ordering.
- Read [troubleshooting](references/troubleshooting.md) for Corepack/pnpm,
  stale OpenAPI, auth redirects, build/type errors, MSW, and seeded E2E data.
- Run `python scripts/frontend_preflight.py --repo <checkout>` for a read-only
  Node/package/layout check. It does not install packages or launch Next.js.

## Development sequence

1. Work from `autogpt_platform/frontend`. Enable Corepack so the
   `packageManager` field selects the pinned pnpm version, then install only
   the dependencies required for the task.
2. For a page, use the App Router under `src/app`, keep page logic in a hook
   when it is non-trivial, and colocate feature components and helpers.
3. For server data, use Orval-generated React Query hooks under
   `src/app/api/__generated__/endpoints/`; regenerate from the backend OpenAPI
   source rather than hand-editing generated files.
4. Use Tailwind design tokens, the design-system components, and the Icon atom
   with Hugeicons. Do not introduce legacy component paths or raw internal
   links when the project convention provides a component.
5. Add a page-level Vitest/RTL/MSW integration test for new feature behavior.
   Use Playwright only for a critical real-browser journey and Storybook for
   visual/design-system behavior.

## Required checks after frontend changes

```bash
pnpm format
pnpm lint
pnpm types
pnpm test:unit
```

Use `pnpm generate:api` after intentional backend API/OpenAPI changes. Full
`pnpm test` or `pnpm test-ui` builds the app and requires the backend stack and
seeded test data; do not run it as a casual smoke check.

## Routing boundaries

- Backend route/schema/OpenAPI source changes belong to
  [platform-backend](../platform-backend/SKILL.md).
- Docker, service startup, env files, and migrations belong to
  [platform-stack](../platform-stack/SKILL.md).
- Classic frontend/agent references belong to [classic-agents](../classic-agents/SKILL.md).

Files in this skill

  • SKILL.md2.8 KB
  • references/api-and-testing.md3 KB
  • references/frontend-architecture.md3.7 KB
  • references/troubleshooting.md2.5 KB
  • scripts/frontend_preflight.py3.7 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.