skilld
All skills
vercel-labs avatar

/dogfood

@8c15ff9 official
by Vercel Labsvercel-labs/agent-browser43k stars
2,928

Systematically explore and test a web application to find bugs, UX issues, and other problems. Use when asked to "dogfood", "QA", "exploratory test", "find issues", "bug hunt", "test this app/site/platform", or review the quality of a web application. Produces a structured report with full reproduction evidence -- step-by-step screenshots, repro videos, and detailed repro steps for every issue -- so findings can be handed directly to the responsible teams.

  • 3 files
  • 15.6 KB
  • Updated 3 weeks ago
  • GitHub

Ask your Agent

Use this Skill: https://skilld.dev/gh/vercel-labs/agent-browser/dogfood

This session only. Nothing lands on disk.

templatesdogfood-report-template.md

≈368 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Dogfood Report: {APP_NAME}

Field Value
Date {DATE}
App URL {URL}
Session {SESSION_NAME}
Scope {SCOPE}

Summary

Severity Count
Critical 0
High 0
Medium 0
Low 0
Total 0

Issues

<!-- Copy this block for each issue found. Interactive issues need video + step-by-step screenshots. Static issues (typos, visual glitches) only need a single screenshot -- set Repro Video to N/A. -->

ISSUE-001: {Short title}

Field Value
Severity critical / high / medium / low
Category visual / functional / ux / content / performance / console / accessibility
URL {page URL where issue was found}
Repro Video {path to video, or N/A for static issues}

Description

{What is wrong, what was expected, and what actually happened.}

Repro Steps

<!-- Each step has a screenshot. A reader should be able to follow along visually. -->

  1. Navigate to {URL} Step 1

  2. {Action -- e.g., click "Settings" in the sidebar} Step 2

  3. {Action -- e.g., type "test" in the search field and press Enter} Step 3

  4. Observe: {what goes wrong -- e.g., the page shows a blank white screen instead of search results} Result

Source: SKILL.md on GitHub

Third-party checks

No third-party reports yet.

Provenance

Signed by skilld at 8c15ff9. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 5 hours ago.

Activeupdated 3 weeks ago

Capability

What it can do
Runs commands
All 2 allowed tools
Bash(agent-browser:*)Bash(npx agent-browser:*)

Topics

  • Testing
  • qa
  • browser-automation
  • exploratory-testing
  • bug-hunting
  • agent-browser
  • web-testing

README badge

README badge for vercel-labs/agent-browser/dogfood

What it does

Systematically explores a web application to identify bugs, UX issues, and other problems, then produces a structured report with step-by-step screenshots, repro videos, and detailed reproduction steps for each finding. Uses the agent-browser tool to navigate, interact with the app, and document issues with full evidence so teams can reproduce and fix them.

Generated from the current SKILL.md.

Frequently asked

What is the difference between a static issue and an interactive issue?
Static issues (typos, placeholder text, misalignment, console errors on load) are visible without user interaction and only need a single annotated screenshot. Interactive issues (functional bugs, UX problems, console errors triggered by actions) require user interaction to reproduce and must include a repro video plus step-by-step screenshots.
Do I need to ask for clarification before starting the dogfood session?
No. If the user provides a target URL, start immediately with defaults for session name, output directory, scope, and authentication. Only ask clarifying questions if authentication is mentioned but credentials are missing.
Should I use `npx agent-browser` or `agent-browser` directly?
Always use `agent-browser` directly. The direct binary uses a fast Rust client, while `npx` routes through Node.js and is significantly slower.
When should I document issues — as I find them or after exploring the whole app?
Document each issue immediately as you find it before moving on. Append to the report right away so findings are preserved if the session is interrupted. Never batch all issues for the end.
How many issues should I aim to find before wrapping up?
Aim for 5-10 well-documented issues. Depth and quality of evidence matter more than total count — 5 issues with full repro is better than 20 with vague descriptions.

Generated from the current SKILL.md. These answers refresh after source changes.