2.5k
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill openclaw/skills --skill create- SKILL.md4.5 KB
Overview
This skill guides Claude through creating new browser automation scripts using the stagehand CLI and the Browserbase function model. It explains the interactive exploration, selector identification, local testing, and deployment steps needed to build reliable automations. The focus is practical: inspect the site, verify interactions, scaffold a function, and publish.
How this skill works
The skill walks you from initial discovery to a runnable function. Start a local session to inspect the DOM and take snapshots, identify robust selectors and wait conditions, and test each interaction with stagehand commands. Scaffold a function project, add Playwright-based logic, run local invokes for validation, and publish to Browserbase when ready.
When to use it
- You need to automate repetitive website tasks (form submission, navigation).
- You want to scrape or extract structured data from web pages.
- You are building a Browserbase Function to run automations in production.
- You must debug network requests or API calls used by the site.
- You are starting a new automation project from scratch.
Best practices
- Prefer stable selectors such as data attributes (data-testid) and avoid brittle nth-child rules.
- Always wait for networkidle, navigation, or use waitForSelector for dynamic content.
- Wrap risky steps in try/catch and return structured error objects for observability.
- Capture network logs when API interaction matters and validate responses before processing.
- Test locally with stagehand/BB dev and perform real invokes before publishing.
Example use cases
- Create a price monitor that fetches a product page, extracts price text, and returns a numeric value.
- Automate login and scheduled data extraction for a private dashboard requiring authentication.
- Scrape paginated search results and normalize fields into JSON for downstream processing.
- Build a function that fills a multi-step form and verifies confirmation messages.
- Debug third-party API calls by enabling network capture and replaying requests.
FAQ
Prefer data attributes or semantic IDs; fallback to visible text. Avoid positional selectors and rely on accessible roles when possible.
Can I debug API calls the site makes?
Yes—enable stagehand network capture to list and inspect requests and responses while interacting with the site.