dogfood

Explore web applications via automated browser sessions and document bugs with evidence.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/andreasasprou/polaris --skill dogfood-andreasasprou
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dogfood
Source: https://github.com/andreasasprou/polaris/tree/main/.agents/skills/dogfood
Command: npx skills add https://github.com/andreasasprou/polaris --skill dogfood-andreasasprou

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill finds, verifies, and documents functional, UX, visual, console, and performance issues in web applications so teams receive reproducible evidence they can act on immediately.

Core Features & Use Cases

  • Systematic exploratory testing: Drive automated browser sessions to traverse navigation, exercise forms, and stress interactive flows while checking edge cases and error handling.
  • Repro-first evidence capture: Record step-by-step screenshots and screen recordings for interactive bugs, annotated images for static issues, and save authentication state for repeatable repros.
  • Structured reporting & handoff: Create a templated Markdown report with indexed ISSUE blocks, severity counts, and attachments suitable for attaching to GitHub PRs or bug trackers.
  • Use Case: Run a focused session against a staging billing page to reproduce a payment flow error, capture a repro video, and deliver an ISSUE block ready for engineering.

Quick Start

Start a named dogfood session against the target URL and let the agent explore, record evidence, and produce the report in the configured output directory.

Frequently Asked Questions about dogfood

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I capture reproducible evidence for web application bugs during exploratory testing?

Exploratory testing for web application bugs can capture reproducible evidence by driving automated browser sessions that record step-by-step screenshots, screen recordings, and saved authentication state into a structured markdown report. This ensures interactive bugs are fully documented for immediate engineering handoff.

Can I reproduce intermittent bugs across staging and local environments using automated browser sessions?

Automated browser sessions can reproduce intermittent bugs across staging and local environments by saving authentication state and applying snapshot commands. This enables repeatable repros against specific pages like a billing flow, verifying the failure consistently before generating an indexed report.

What is the best way to document UX issues and console errors found during functional web testing?

Documenting UX issues and console errors during functional web testing works best by capturing annotated screenshots for static issues and screen recordings for interactive flows. A templated markdown report with indexed ISSUE blocks and severity counts organizes the evidence for bug trackers.

Do I need an agent-browser session to run visual testing on authentication flows?

Visual testing on authentication flows requires an agent-browser session to traverse navigation, exercise forms, and stress interactive edge cases. The browser session captures screenshots and video while saving the authentication state, which is necessary for producing repeatable step-by-step repro evidence.

Why should I use automated exploratory testing instead of manual QA for reproducing payment flow errors?

Automated exploratory testing for reproducing payment flow errors provides structured reproducible evidence that manual QA lacks. It drives browser sessions to stress interactive flows, captures video recordings of the exact failure, and outputs a templated ISSUE block ready to attach to a GitHub PR.

What limitations exist when capturing screenshots and video recordings for web app bug reports?

Limitations when capturing screenshots and video recordings for web app bug reports depend on the agent-browser session's ability to interact with the target environment. Complex interactive flows may require saved authentication state and specific snapshot commands to ensure the reproducible evidence accurately reflects the original failure.