Skip to content

AGENTS.md

Latest commit

 

History

History
43 lines (27 loc) · 1.94 KB

AGENTS.md

File metadata and controls

43 lines (27 loc) · 1.94 KB
Raw

AGENTS.md

Cursor Cloud specific instructions

Overview

Video Driven Skill is a two-service app (Java backend + React frontend) that transforms screen recordings into runnable automation skill packages.

Services

Service Directory Dev command Port
Backend (Spring Boot) backend/ ./mvnw spring-boot:run -DskipTests 8080
Frontend (Vite + React) frontend/ npm run dev 3000

The Vite dev server proxies /api and /ws to the backend at localhost:8080 (configured in frontend/vite.config.js).

Startup order

Start the backend first — the frontend proxies API calls to it. The backend creates its SQLite DB and data directories automatically on first run at ~/video-driven-skill/.

Lint & format

  • Backend: cd backend && ./mvnw spotless:check (Google Java Format). Auto-fix with ./mvnw spotless:apply.
  • Frontend: No dedicated lint script configured; use npm run build to catch TypeScript/build errors.

Tests

  • Backend: cd backend && ./mvnw test (no test sources currently present).
  • Frontend: No test framework configured.

Build

  • Backend: cd backend && ./mvnw -DskipTests package produces a fat JAR in target/.
  • Frontend: cd frontend && npm run build produces static assets in dist/.

Key gotchas

  • The Maven enforcer plugin requires Java 17+. The environment has Java 21, which is compatible.
  • The backend uses Spring Boot 4.1 RC1 with the Spring milestone repository configured in the Maven wrapper properties. First-time dependency resolution takes ~20s.
  • FFmpeg must be on PATH for video frame extraction. The FFMPEG_PATH env var can override this.
  • AI generation features require AI_API_KEY, AI_BASE_URL, and AI_MODEL env vars (or .env file). The app still works without them for upload/extract/edit/run workflows.
  • The backend runner.timeout defaults to 600 seconds for skill execution.