Skill
webapp-testing Skill: drive local web apps with Playwright
Official Anthropic webapp-testing Skill: Python Playwright scripts for local web apps, not tested
This is a skill folder in the anthropics/skills repository. The page follows the SKILL.md text and is not tested.
- Verified
- · Not tested by us
- Applies to
- How to invoke
- No official command
No verbatim invocation command for this skill was found in this check; after installing the whole example-skills plugin, the README says to mention the skill in conversation (its example is the PDF skill, so for this skill it is by analogy); SKILL.md names it webapp-testing[1][4]
- Install
/plugin marketplace add anthropics/skills /plugin install example-skills@anthropic-agent-skillsInstalls the whole example-skills plugin (webapp-testing plus 11 other skills), not this skill alone; a single-skill install is not stated officially[4][5]
- License
- Apache-2.0 [2]
- Prerequisites
- Checked revision
dbd4588- Tools
When to use
- You need to verify frontend functionality of a local web app or debug UI behavior; the SKILL.md description lists this.[1]
- You need browser screenshots or browser logs; the description lists these too.[1]
When not to use
- The target is not a local web app (inference): SKILL.md is written for local apps such as localhost servers; other cases are not stated officially.[1]
- You do not want to use Python scripts (inference): SKILL.md asks for native Python Playwright scripts and offers no other route; the official text does not list this as a non-use case.[1]
Purpose and approach
The SKILL.md description says this is a toolkit for interacting with and checking local web applications using Playwright, supporting frontend verification, UI debugging, browser screenshots and viewing browser logs. The body asks you to write native Python Playwright scripts.[1]
- A bundled script, scripts/with_server.py, manages the lifecycle of one or more servers
- SKILL.md says to run scripts with --help first and only read the source if truly necessary
- The examples/ folder has three examples: element discovery, static HTML automation, console logging
The official decision flow
| Situation | What to do |
|---|---|
| Static HTML | Read the HTML to find selectors, then write a Playwright script; if that fails, treat it as dynamic |
| Dynamic app, server not running | Run python scripts/with_server.py --help first, then use it with a simplified script |
| Dynamic app, server already running | Reconnaissance then action: wait for networkidle, screenshot or inspect the DOM, then act with the selectors found |
Pitfall called out officially: on dynamic apps, wait for networkidle before inspecting the DOM. The SKILL.md example launches chromium in headless mode.[1]
Dependencies, install and license
| Item | What the official text says |
|---|---|
| Runtime | Python Playwright scripts; how to install Playwright is not stated officially |
| Install command | No install command for this skill on its own; the root README gives the install command for the whole example-skills plugin (below) |
| License | Apache License 2.0 (from the LICENSE.txt text); the frontmatter says see LICENSE.txt |
| Account, permissions, cost | Not stated officially |
The repository root README at the same SHA says to run /plugin marketplace add anthropics/skills to register the repo as a marketplace, then /plugin install example-skills@anthropic-agent-skills; the example-skills list in marketplace.json includes ./skills/webapp-testing. This installs the whole plugin (with 11 other skills) rather than webapp-testing alone; a single-skill install is not stated officially. After installing, the README says you can use a skill just by mentioning it; its example is the PDF skill, so applying that to webapp-testing is by analogy.[4][5][3]
Sources
All links are official pages; verification dates are listed per source, and links point to the version checked.