Explore a web app like a user. Find issues. Prove every finding.
When to use
- User asks to dogfood, QA, exploratory test, bug hunt, or test a site/app.
- Target is a web app or local URL.
- User wants actionable issues with repro evidence.
Inputs
- Required: target URL.
- Optional: session name, output directory, scope, credentials.
- Defaults:
- Session: slugified domain.
- Output:
./dogfood-output/. - Scope: full app.
- Auth: none.
If target URL is present, start. Ask only when auth is needed and credentials are missing.
Rules
- Use
agent-browserdirectly. - Do not use
npx agent-browser. - Do not read target app source code.
- Do not delete output files.
- Do not close and restart session mid-run.
- Verify issue before documenting.
- Document each issue immediately.
- Static issue: annotated screenshot enough.
- Interactive issue: video plus step screenshots.
Setup
mkdir -p {OUTPUT_DIR}/screenshots {OUTPUT_DIR}/videos
cp {SKILL_DIR}/templates/dogfood-report-template.md {OUTPUT_DIR}/report.md
agent-browser --session {SESSION} open {TARGET_URL}
agent-browser --session {SESSION} wait --load networkidle
If login is needed:
agent-browser --session {SESSION} snapshot -i
agent-browser --session {SESSION} fill @e1 "{EMAIL}"
agent-browser --session {SESSION} fill @e2 "{PASSWORD}"
agent-browser --session {SESSION} click @e3
agent-browser --session {SESSION} wait --load networkidle
agent-browser --session {SESSION} state save {OUTPUT_DIR}/auth-state.json
Ask user for OTP codes.
Flow
- Create output directories and report.
- Open target and authenticate if needed.
- Take initial annotated screenshot.
- Read
references/issue-taxonomy.md. - Explore top-level navigation.
- Test core workflows, forms, modals, menus, empty states, and errors.
- Check console and browser errors.
- Document each issue as soon as found.
- Stop around 5-10 well-documented issues.
- Update severity counts and close session.
Explore commands
agent-browser --session {SESSION} screenshot --annotate {OUTPUT_DIR}/screenshots/initial.png
agent-browser --session {SESSION} snapshot -i
agent-browser --session {SESSION} errors
agent-browser --session {SESSION} console
Use snapshot -i for interactive refs.
Use snapshot for text and content.
Evidence
For interactive issues:
agent-browser --session {SESSION} record start {OUTPUT_DIR}/videos/issue-{NNN}-repro.webm
agent-browser --session {SESSION} screenshot {OUTPUT_DIR}/screenshots/issue-{NNN}-step-1.png
sleep 1
agent-browser --session {SESSION} screenshot --annotate {OUTPUT_DIR}/screenshots/issue-{NNN}-result.png
sleep 2
agent-browser --session {SESSION} record stop
For static issues:
agent-browser --session {SESSION} screenshot --annotate {OUTPUT_DIR}/screenshots/issue-{NNN}.png
Output
- Report path.
- Total issues.
- Severity breakdown.
- Most critical findings.
- Evidence files saved.