Agent Skills for Deployment and Browser Work

Shipping and checking a web app involves different permissions. Separate preview deployment, token handling, optimization audits, and browser-driven verification before choosing a Skill.

These instructions do not grant Vercel access, supply tokens, or install a browser CLI. Confirm the project, account, and approval boundary before any operation that changes remote state.

Selected Skills and Original Sources

Deploy to Vercel

Prepare a Vercel deployment by checking project linkage, CLI access, and team context before choosing an authorized deployment route.

Read Local Guide →Publisher: Vercel Labs

Vercel CLI with Tokens

Use Vercel CLI token workflows with explicit project context and careful handling of credentials and organization identifiers.

Read Local Guide →Publisher: Vercel Labs

Vercel Optimize

Investigate Vercel performance using observability signals, scoped checks, and prioritized recommendations grounded in available project data.

Read Local Guide →Publisher: Vercel Labs

agent-browser Skill

Guide an agent through the agent-browser CLI, using tool-version-specific instructions for browser automation tasks and specialized workflows.

Read Local Guide →Publisher: Vercel Labs

Webapp Testing

Uses Playwright to inspect local web apps, verify frontend behavior, capture screenshots, and examine browser logs.

Publisher: Anthropic

This is an independent directory. Check the publisher's source for installation, tools, and availability.

Choose by the Job

A Preview Needs Deployment

Deploy to Vercel is about checking project linkage and creating a reviewable deployment.

A Token Is Involved

CLI with Tokens focuses on authentication boundaries and avoiding accidental exposure of secrets.

The Running Site Needs Checking

Use browser or webapp testing guidance to inspect actual behavior; use Optimize for an evidence-led performance and cost audit.

Explore Other Skill Categories

Choose the next task, not just the next publisher.

Frequently Asked Questions

How do deployment instructions differ from browser automation?
Deployment changes where or how an application is served. Browser automation operates a browser to inspect or carry out a task. They may be used together, but neither one proves that a deployed application actually works for the user.
What should I confirm before deploying to Vercel?
Confirm the repository, project, account or team, target environment, and the approval boundary. Review the relevant build and environment settings so a preview operation does not accidentally become a production change.
Does a successful HTTP response prove a page is usable?
No. A server can return 200 while a client-side request fails and the page shows an error or no content. Inspect the rendered page, expected text, and key interaction rather than treating the response code as a content check.
Is agent-browser available just because its Skill is installed?
No. The CLI and its runtime prerequisites are separate from the instruction file. Check the tool’s installation and supported commands, then use guidance that matches that version.
How should tokens be handled in a CLI workflow?
Use the approved credential source and the narrow permissions required for the task. Avoid placing secrets in command arguments, shared logs, screenshots, or published files, and confirm which project or organization a command will affect.
What should I check after a deployment finishes?
Open the public target and verify representative routes, assets, expected content, and a meaningful user action. Record the actual deployment version; a successful build or preview alone does not establish production state.

Keep Deployments and Checks Traceable

Review the target, permissions, browser evidence, and next action in one workspace.