Skip to main content
plugin CX Codex

assay — Claude Code plugin: page checks and setup

assay checks changed web pages by interacting with their controls, then reports the results at the end of the agent’s turn. It helps catch failures after a page edit.

MIT ★ 1,596 updated 2026-10-29

Install into Claude Code · Plugin marketplace All methods and platforms ↓
/plugin marketplace add hashgraph-online/awesome-codex-plugins
/plugin install assay@awesome-codex-plugins
Open on GitHub

What assay does

When a turn changes a web page, the plugin runs a check and reports a clean result or lists findings. It exercises page controls and links repeated errors to the first finding rather than restating it. The plugin reports problems but does not edit code unless you ask.

A Stop hook starts the check after a page-changing turn. A changed web file marks an HTML page in the same folder; when timestamps tie, index.html takes precedence. The search goes no deeper than four folders, skips node_modules and .git, and checks up to three pages per turn. Each page gets up to four minutes, with up to nine minutes for the complete run.

In Claude Code, installation includes both the skill and the hook. A checked page always produces a report, even when no issues appear; turns without page changes stay quiet. assay cannot tell whether a page behavior is intentional, so a control that does nothing in its current state is not automatically a confirmed bug.

What it helps with

  • check controls on a recently edited page
  • find errors triggered by clicking buttons
  • group repeated failures under one finding
  • see which pages were checked after an edit

Example prompts for the agent

  • Edit app.js and check the page in the same folder.
  • Check every web page changed in this turn and report the results.
  • After the edits, test what happens when each button is clicked.

How to install assay

Plugin marketplace

/plugin marketplace add hashgraph-online/awesome-codex-plugins
/plugin install assay@awesome-codex-plugins

Step by step, with commands for each platform: Claude Code plugins: marketplaces, installing, and building your own

Questions about assay

How do I install assay in Claude Code?

Run these commands in order: /plugin marketplace add hashgraph-online/awesome-codex-plugins, then /plugin install assay@awesome-codex-plugins.

What does assay do?

After a page-changing turn, the plugin tests the page controls and reports what happened. It leaves code changes to the agent unless you request a fix.

How does assay choose pages to check?

A changed web file marks an HTML page in its folder. The search covers up to four folder levels, skips node_modules and .git, and checks no more than three pages per turn.

Source text

assay · MIT
<h1 align="center">The Skill and the Plugin</h1>

<p align="center"><b>English</b> · <a href="README.zh-CN.md">简体中文</a> · <a href="README.ja.md">日本語</a> · <a href="README.ko.md">한국어</a> · <a href="README.ru.md">Русский</a></p>

---

<p align="center">Two ways to have your agent check a page it just wrote.
<strong><a href="../../docs/harnesses.md">Install steps for fifteen harnesses →</a></strong></p>

---

## Contents

- [The Skill](#the-skill)
- [The Plugin](#the-plugin)
- [Reporting, Not Fixing](#reporting-not-fixing)

---

## The Skill

The skill works in all fifteen supported harnesses.
[`skills/checking-a-page/SKILL.md`](skills/checking-a-page/SKILL.md) tells the
agent to run this after finishing a page:

```bash
assay <the page> --one-line
```

and to end its reply with the output, word for word:

```
assay: checked todo/todo.html, 8 checks, nothing flagged.
```

or, when something is wrong:

```
assay: checked notes/index.html, 12 checks, 4 flagged:
  - press Clear all: the page threw an error and stopped running, Cannot set properties of null (setting 'innerHTML')
  - press Clear all twice (same finding as press Clear all)
  - press Add, then Clear all (same finding as press Clear all)
  - press Clear all, then Add (same finding as press Clear all)
```

Those four lines are one bug. The repeats name the first finding instead of
repeating its message. The agent copies assay's output rather than
summarising it, so the numbers and wording stay exact. With the skill alone,
the agent decides when to run the check.

---

## The Plugin

The plugin adds a `Stop` hook, which runs the check at the end of every turn
that changed a page. In Claude Code, Codex, Grok Build CLI, Devin CLI, Factory
Droid and GitHub Copilot CLI it installs the skill and the hook together. In
Cursor it installs the skill alone. The DeepSeek Harness runs the hook through
its Claude Code bridge.

- **Which pages it checks.** A change to any web file marks the page in the
  same folder, so editing `app.js` counts. In each folder it checks the most
  recently changed `.html` file, preferring `index.html` on a tie.
- **How far it looks.** It searches at most four folders deep, skips
  `node_modules` and `.git`, only considers files changed since its last run,
  and checks at most three pages per turn.
- **It always reports** when it checks a page, so a clean result can't be
  mistaken for assay not running. A turn that changed no page reports
  nothing.
- **It keeps the turn open** so you see the result, and never does this twice
  in a row. It checks the harness's `stop_hook_active` flag and writes a
  timestamp (in the harness's scratch folder, else the repository's `.git`
  folder, else a `.assay-last-run` file). If it can't write the timestamp, it
  reports without keeping the turn open.
- **It finishes in time.** It gives each page up to four minutes and all
  the pages together up to nine, so it ends before the harness cancels it at
  the ten-minute limit set in `hooks.json`. A page left without time is
  reported as not checked.

---

## Reporting, Not Fixing

The agent shows you what assay found and doesn't change code because of it
unless you ask. assay doesn't know what the page is for, so a finding is a
place to look, not a confirmed bug. A control can correctly do nothing in the
state it was pressed in, and a value can be limited on purpose. Ask the agent
to fix something and it will.

© awss1i · MIT · hashgraph-online/awesome-codex-plugins

Similar plugins

Plugins catalog →
Assay MCP Server Rul1an Assay MCP Server is an MCP server for Claude Code that gates tool calls by policy and records decisions in offline-verifiable evidence bundles. CX Codex +1 ★ 16repo stars ub-mode unclebob The ub-mode skill for Claude Code helps developers structure code changes. It guides the development process by following the principles of "First Make it Work, Then Make it Right, Then Follow Practice." CX Codex +1 930installs methodology-advisor FlorianBruniaux The methodology-advisor command for Claude Code and Codex reads the codebase and asks three questions to recommend an AI-development methodology stack. Run it at the start of a project, before choosing a workflow. CX Codex +1 ★ 6,529repo stars discernment-nudge official anthropics discernment-nudge — a plugin for Claude Code and Claude.ai that appends two or three short questions after a substantive answer. Helps check key facts, probe the reasoning, and spot missing context. ★ 198,522repo stars academy-guide official anthropics academy-guide is a Claude Code plugin that points users to Claude Academy courses, tutorials, and use cases. It responds to questions about using Claude. ★ 198,522repo stars claude-api official anthropics This plugin adds Claude API and SDK guidance for building LLM-powered applications. Install it when your project needs a Claude integration. ★ 198,522repo stars example-skills official anthropics example-skills — a plugin for Claude Code and Claude.ai with a set of sample skills. Demonstrates how to build skills, MCP servers, artifacts, web tests, and internal communication tools. ★ 198,522repo stars document-skills official anthropics Document processing suite for Excel, Word, PowerPoint, and PDF files. Equips the assistant to inspect and generate common document formats. ★ 198,522repo stars

From the same set391

The whole set →
babok_analyst babok_analyst guides business analysis across 9 BABOK v3 stages with mandatory human sign-off at every gate. CX Codex ★ 1,596repo stars aw1-circuit-breaker The aw1-circuit-breaker plugin checks ASTs and tool-call arguments before Python runs code. Use it to intercept dynamic execution, shell launches, and socket connections. CX Codex ★ 1,596repo stars auto-pilot auto-pilot turns a short Claude request into a one-question-at-a-time intake, saves the answers as project files, and lets a later session pick up where the last one stopped. CX Codex ★ 1,596repo stars atomlane atomlane compiles safe parallel execution plans for builds, test suites, and agent scripts. It runs tasks concurrently only when dependencies permit. CX Codex ★ 1,596repo stars armorcodex armorcodex is a Claude Code plugin that checks Bash, apply_patch and MCP tool calls against a declared intent plan and policy rules. CX Codex ★ 1,596repo stars repobrain repobrain builds a knowledge base for a repository and answers code questions with file and line references. Use it to explore an unfamiliar project or trace how its modules connect. CX Codex ★ 1,596repo stars anchor anchor — a discipline pack for Claude and Claude CLI: a skill, 22 slash commands and 5 hooks that keep the agent inside the task and carry project memory across sessions. CX Codex ★ 1,596repo stars alcove The alcove plugin for Claude Code and Codex is a local documentation server with hybrid BM25 and vector search. It retrieves relevant project context on demand. CX Codex ★ 1,596repo stars

Project sponsors

Companies that keep this analytics open