Skip to content
Get packs

Dogfood Branch QA

Browser-tests your current branch's changes end to end, fixes small bugs, and writes a QA report

Source Download skill
Repository
EveryInc/compound-engineering-plugin
Stars
25,850
Licence
MIT
Files
7 files

What it does

When you invoke it, the agent acts like a QA engineer: it maps the user flows affected by your branch's changes versus the trunk, drives a real browser through them, and judges both correctness and how the experience feels to your product's personas. It can make small, low-risk fixes, add regression tests, and commit them, while escalating bigger decisions to a human. The result is a durable markdown report saved in your repo's dogfood-reports folder.

Great for

  • Developers who want an automated end-to-end browser check of a feature branch before opening a pull request
  • Teams using Compound Engineering packs that define personas and UX criteria
  • People who want a written QA report of what was tested and what remains for a human

Skip it if

  • Whole-app exploratory testing, since it is scoped to the branch diff by design
  • Repos where the agent should not auto-commit changes, because it commits fixes itself (risk of unreviewed commits)
  • Projects without a runnable local dev server, since the run cannot serve the app

Try asking

  • “/ce-dogfood”
  • “/ce-dogfood 123 --port 3000”
  • “/ce-dogfood feature/checkout-redesign”
Step-by-step guide: how to use Dogfood Branch QAInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 7 original files, the MIT licence and an attribution note.

    ce-dogfood.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip ce-dogfood.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip ce-dogfood.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Invoke with no argument on the branch you want tested, or pass a PR number so it is isolated in a worktree without touching your checkout
  • Review the Decisions for a human section and any auto-commits in the report before merging
  • Run it on a clean working tree so auto-fix commits are easy to inspect and revert

What's inside

  • agents/openai.yaml
  • references/agent-browser-driver.md
  • references/dogfood-report-template.md
  • references/phases.md
  • references/test-matrix-taxonomy.md
  • scripts/packs-resolve.py
  • SKILL.md

Redistributed unmodified from EveryInc/compound-engineering-plugin under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.