frontend-dev
/frontend-dev
Manages local web server lifecycle and enables web page development with Playwright MCP. Use when developing HTML/CSS/JS, building web pages, viewing web pages, testing UI interactions, verifying renders, debugging frontend issues, or iterating on web design. Covers build, startup, shutdown, and browser-based testing workflows.
- Author
- j-r-beckett ↗
- Repository
- SpeedReader ↗
- Licence
- Apache-2.0
- Last repository change
- 22.05.2026
Assessment and security
- Files in this package
- 7
- Scripts
- 6
- HTML comments
- 0
- URLs in SKILL.md
- 3
No signals detected. Repository changed on 22.05.2026, without references to retired models or APIs.
Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.
- 6 scripts in this package. Review them before installing, as Claude may execute them.
Scripts: scripts/actions/demo_upload.js, scripts/actions/toggle_layers.js, scripts/reload.py, scripts/resolve_action.py, scripts/setup.py, scripts/watchdog.py
What it does and when to use it
Manages local web server lifecycle and enables web page development with Playwright MCP. Use when developing HTML/CSS/JS, building web pages, viewing web pages, testing UI interactions, verifying renders, debugging frontend issues, or iterating on web design. Covers build, startup, shutdown, and browser-based testing workflows.
Original author description
Manages local web server lifecycle and enables web page development with Playwright MCP. Use when developing HTML/CSS/JS, building web pages, viewing web pages, testing UI interactions, verifying renders, debugging frontend issues, or iterating on web design. Covers build, startup, shutdown, and browser-based testing workflows.
How it works
- Setup
- Development Loop
- Actions
- Fixes
SKILL.md preview
---
name: frontend-dev
description: Manages local web server lifecycle and enables web page development with Playwright MCP. Use when developing HTML/CSS/JS, building web pages, viewing web pages, testing UI interactions, verifying renders, debugging frontend issues, or iterating on web design. Covers build, startup, shutdown, and browser-based testing workflows.
---
# Frontend Development
Toolkit for frontend development in the SpeedReader codebase. Handles server lifecycle and delegates browser-based testing to Playwright MCP.
## Setup
Run setup first:
```bash
uv run .claude/skills/frontend-dev/scripts/setup.py
```
Outputs:
- `READY` - Proceed with usage
- `RESTART_REQUIRED` - Ask user to restart Claude Code, then stop and wait
## Development Loop
**All scripts must be run with `uv run`, not `python3`.**
### 1. Edit code
Make changes to frontend code in `src/Frontend/`.
### 2. Reload
```bash
uv run .claude/skills/frontend-dev/scripts/reload.py
```
This builds SpeedReader, starts the server, and waits for health. The server runs in the background with a 15-minute auto-shutdown.
`reload.py` is **fast**. Claude should *never* attempt to short-circuit it.
`reload.py` does a complete, fresh build of the frontend. It is *not necessary* to run **any** build steps manually.
**SpeedReader does not hot reload. Run reload.py after every code change.**
### 3. Interact with Playwright MCP
Use Playwright MCP tools to interact with the page at http://localhost:5050/demo.
For visual regression testing, use `?hideProcessingTime=true` to hide dynamic content:
```
http://localhost:5050/demo?hideProcessingTime=true
```
**Snapshot vs Screenshot:**
- Use `browser_snapshot` for interaction (getting element refs) and existence checks
- Use `browser_take_screenshot` for visual verification - things that look fine in a snapshot are often revealed to be broken when viewed as a screenshot
As a general rule, use snapshots for interaction and screenshots for verification and validation.
*Important*: when looking at screenshots, *be careful*. It's very easy to say "this is fine" when the screenshot is close to the expected result, but different in a visually similar but semantically different way. Example: a rectangle with pointed corners vs rounded corners, small adjustments to the position of an item, the levels of indentation in a list. For cases like this, identify the area of interest, look very carefully at it, and try to identify physically small yet semantically distinctive features.
### 4. Repeat
Edit code -> reload -> verify with Playwright -> repeat until done.
## Actions
Actions are reusable JS scripts. Call them via `browser_run_code`:
```
browser_run_code: code="@demo_upload"
browser_run_code: code="@demo_upload?file=test.png"
browser_run_code: code="@toggle_layers?a_rect=true;r_rect=false;polygon=false;text=true"
```
Syntax: `@action` or `@action?key=value;key2=value2`
Available actions:
- `demo_upload` - Upload image and wait for OCR. Args: `file`
- `toggle_layers` - Set visualization layer visibility. Args: `a_rect`, `r_rect`, `polygon`, `text` (all optional, true/false)
### Writing Actions
Actions are JavaScript files in `scripts/actions/`. Each is an async function receiving `page` and `args`:
```javascript
async (page, args) => {
const file = args.file || 'hello.png';
await page.locator('#file-input').setInputFiles(file);
await page.waitForSelector('.stats-bar', { timeout: 60000 });
}
```
Create or edit actions as needed.
## Fixes
### Screenshots timing out
If `browser_take_screenshot` hangs or times out repeatedly, close and reopen the browser:
```
browser_close
browser_navigate: url="http://localhost:5050/demo"
```
This resets the browser state and typically resolves the issue.
Files in this package (7)
- SKILL.md 4 KB
- scripts/reload.py 6 KB
- scripts/resolve_action.py 3 KB
- scripts/setup.py 2 KB
- scripts/watchdog.py 1 KB
- scripts/actions/demo_upload.js 1 KB
- scripts/actions/toggle_layers.js 1 KB
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · j-r-beckett/SpeedReader | 100 | Apache-2.0 |
| unicomai/wanwu | 95 | Apache-2.0 |
| MiniMax-AI/skills | 95 | MIT |
| marmelab/atomic-crm | 90 | MIT |
Database snapshot: 13.11.2026 · package version a535cc85cdb2 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100webapp-testing
Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs. Use when testing web apps, automating browser interactions, or debugging frontend issues. - 100verdict
Token-efficient browser verification for AI coding agents. Use instead of Playwright MCP for all browser testing — page verification, click flows, form fills, CSS inspection, screenshots. Costs ~50-100 tokens per call vs ~1500 for MCP tools. Use when verifying web pages,… - 100web-debug
Python Playwright testing of local web apps. Use when verifying frontend functionality, debugging UI behavior, capturing browser screenshots, or viewing browser console logs. - 100web-debug
You MUST use this when interacting with or testing local web applications with Playwright - verifying frontend functionality, debugging UI behavior, capturing browser screenshots, or viewing browser console logs. - 100local-frontend-testing-debugging-150c3767
Use when testing, debugging, or making targeted improvements to rendered frontend apps through the Build Web Apps or web dev plugin: local dev servers, UI regressions, interaction bugs, console errors, responsive layout, and visual QA. Check whether the Browser plugin is… - 100playwright
Playwright MCP server for Claude Code browser automation. Use when configuring Playwright MCP, automating browser interactions, taking screenshots, researching web content blocked by non-browser clients, testing web UI locally, or navigating web pages from Claude Code sessions.