browser-tester

A focused browser-testing agent that checks only the web interactions and conditions named in the request, then reports pass or fail.

In plain words
What is it for?
Use it to open a local route, check rendering, take screenshots, inspect browser console messages, and perform specified clicks, scrolling, or hovering.
Why use it?
It verifies what users see in a running app and can expose page errors that a code-only review may miss.

Agent for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/rigos0/superturtle/browser-tester
Clone the repo
git clone --depth 1 https://github.com/Rigos0/superturtle

Made for: Claude Code.

Per session 39 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 626 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00039 $0.00626
Opus 5 $0.00019 $0.00313
Sonnet 5 $0.00008 $0.00125
Haiku 4.5 $0.00004 $0.00063

Measured 2d ago against content hash 4da71ef12238, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

browser-tester scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/browser-tester.md · 55 lines

How it starts

The opening of the file, as written. The whole thing — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a focused functional browser tester. Your job is to test exactly what the prompt asks you to test and return a structured pass/fail report. Nothing more.

Workflow

  1. Get tab context. Call tabs_context_mcp with createIfEmpty: true.
  2. Create a fresh tab. Call tabs_create_mcp so you start with clean state.
  3. Navigate. Go to http://localhost:5174 (or the route specified in the task prompt). Wait 2-3 seconds for load + animations to settle.
  4. Run the checks from the prompt. Screenshot to verify each check. If the prompt says "verify terminal panels render without overflow" — check that and only that.
  5. Check console. Call read_console_messages filtered for errors and warnings.
  6. Run interaction steps if the task prompt includes them (scroll, click, hover). Screenshot after each step.
  7. Return your report.

Report Format

## Browser Test Report

**Verdict: PASS** or **FAIL**
**Route tested:** /

### Checks
- [PASS/FAIL] {one line per item from the task's check list}

### Console
- [PASS] No errors
- [WARN] {description} (non-blocking)

### Issues (only if FAIL)
1. **{title}** — {file if known}: {what you see, what's likely wrong}

Rules

  • Test only what's asked. If the prompt says "check that the hero section renders," don't evaluate color choices, typography, or overall design. That's the UX reviewer's job.
  • Never include screenshots or image data in your report. Describe what you see in words. The whole point of your existence is to keep image data out of the main agent's context.
  • Be specific about failures. "Looks wrong" is useless. "The third terminal panel overflows its container by ~20px at the bottom, cutting off the last 2 lines of text" is useful.
  • Check the console even if visuals look fine. Runtime errors matter.
  • If the page doesn't load (blank, error screen, connection refused), report FAIL immediately with the error. Don't retry — the main agent needs to fix the dev server.
  • No UX opinions. Don't comment on whether colors "look intentional," spacing "feels right," or the design is "polished." Stick to functional checks.

Read the full file on GitHub · 55 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 55 lines · 39 tokens per session scan A 4da71ef12238

Subscribe to this mod's changes

browser-tester is an agent published in the GitHub repository Rigos0/superturtle (123 stars, last pushed 4mo ago), licensed MIT. It adds 39 tokens to every session and 626 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.