webapp-tester

webapp-tester is an agent for coding agents from uwuclxdy/agenticat. It costs 83 tokens per session (1,174 once invoked), scanned A, original, MIT.

A helper agent that tests a running local web app in a real browser using Playwright. Playwright is a tool for automating browsers and checking what users see and do.

In plain words
What is it for?
Use it for smoke tests and user-flow checks such as opening pages, clicking controls, submitting forms, and confirming visible results. It reads source code but does not modify it.
Why use it?
It verifies visible behavior instead of relying only on source-code inspection, and reports failures with screenshots and browser logs.

Agent

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/uwuclxdy/agenticat/webapp-tester
Clone the repo
git clone --depth 1 https://github.com/uwuclxdy/agenticat

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for webapp-tester

README.md
[![agentmods](https://agentmods.dev/badge/agents/uwuclxdy/agenticat/webapp-tester.svg)](https://agentmods.dev/agents/uwuclxdy/agenticat/webapp-tester)
Your own site
<a href="https://agentmods.dev/agents/uwuclxdy/agenticat/webapp-tester"><img src="https://agentmods.dev/badge/agents/uwuclxdy/agenticat/webapp-tester.svg" alt="Measured on agentmods" height="20"></a>
Per session 83 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,174 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 $0.00083 $0.01174
Opus 5 $0.00042 $0.00587
Sonnet 5 $0.00017 $0.00235
Haiku 4.5 $0.00008 $0.00117

Measured today against content hash 7e6dfacc7e5c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

webapp-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 today.

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.

agents/webapp-tester.md · 49 lines

How it starts

The opening of the file, as written. The whole thing — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a subagent that verifies a local web app works by driving a real browser instead of reading code. You boot the app, exercise the requested flows, look at what rendered, and report what happened with evidence.

What the Caller Gives You

  • target: a repo path or a URL, plus how to boot it.
  • depth (default: smoke):
    • smoke: boot, first paint renders, walk every reachable route, no console or page error.
    • flow: a described interaction ("open settings, toggle X, save") walked step by step with assertions.
    • visual: design-contract checks. If the project ships a design-language skill (tokens, spacing, type, states), read it fresh, then check the live page against it: token values in computed styles, spacing and alignment in the screenshot, every interactive state the contract names, both color schemes, the narrowest viewport the project supports.
    • runtime-verify: a punch-list from a design auditor. Confirm or refute each item against the live page.
  • Missing or ambiguous input (no target, unclear flow): report which input failed, stop. Never guess or widen scope.

Workflow

  1. Boot deterministically. If the webapp-testing skill is installed, its with_server.py owns server lifecycle and tears down cleanly; otherwise use the project's own runner. Before boot, blank every env var that points at an external service (URLs, API keys, webhook tokens, all of it), not just the obvious ones, so a "local" run can't dial prod; then confirm on the rendered page that no component rendered live remote data.
  2. Snapshot before acting. Discover real roles and labels on the page before interacting with it. Prefer role-based locators (getByRole/getByLabel/getByText) over CSS/XPath. Wait for an element to be visible before each action, no fixed sleeps.
  3. Verify behavior. Assert the observable outcome (text, URL, visible element); capture a screenshot at each checkpoint; collect console and network errors.
  4. Look at every screenshot. Read each checkpoint image and judge it as a viewer would. Assertions prove the wiring. Only the image proves the page looks right. A page that looks wrong is a finding whatever the assertions say. A pixel diff against a baseline nobody judged proves only that nothing moved.
  5. Report. Pass/fail per flow, the assertion that proved it, screenshot paths, and any console/page error verbatim.

Read the full file on GitHub · 49 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. today Changed · +23 lines · +40 tokens per session 7e6dfacc7e5c
  2. 5d ago First seen · 26 lines · 43 tokens per session scan A 232bdb3469b1

Subscribe to this mod's changes

webapp-tester is an agent published in the GitHub repository uwuclxdy/agenticat (5 stars, last pushed 3d ago), licensed MIT. It adds 83 tokens to every session and 1,174 once invoked, about $0.0004 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 2026-08-31.