Skip to main content
Glama

Audit a page with pour

pour_audit

Audits a URL or local HTML file against WCAG 2.2 A/AA in headless Chrome, returning each failing rule with severity, success criteria, failing elements as CSS paths, and fixes.

Instructions

Audit a URL against WCAG 2.2 A and AA (or, with wcag3, the requirements of the WCAG 3.0 draft) with the pour engine in headless Chrome: the page is loaded, the engine runs against the live DOM, and every failing rule comes back with its severity, its success criteria, the failing elements as CSS paths with the fix, and the count of rules the engine would not judge without a human. Works on a running dev server or any URL. Results depend on the page state: the viewport width (responsive pages serve different content per breakpoint), whether the page was scrolled first (lazy content), and what is open on it. Say which you used when you report a figure.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "pour_audit"
  }
}

Input Schema

NameRequiredDescriptionDefault
urlNoThe page to load, http or https only. A local dev server is fine (http://localhost:8080/). Without a scheme, https is tried first, then http. For a file on disk use path instead: file:// URLs are refused.
pathNoA file on disk to check instead of a URL, relative to the folder the server was started in. It is served over http from its own folder, so relative assets load. Give one of url or path, not both.
rootNoWith path: the folder to serve as the site root, for a page whose assets are addressed from the root (/styles.css) rather than from its own folder. Defaults to the file's own folder.
wcag3NoAudit against the WCAG 3.0 Working Draft instead of WCAG 2.2 A and AA: core checks its core requirements, supplemental its core and supplemental ones. An unfinished standard: WCAG 2 rules matched to draft requirements, never grounds for a conformance claim. Leave unset unless WCAG 3 is asked for.
detailNosummary: totals only. rules: one entry per failing rule, no elements. elements (default): up to maxNodes elements per rule.
scrollNoScroll through the page first so lazy-loaded content is in the DOM. Changes the numbers, by design. Default false.
waitMsNoExtra settle time after the page loads, in milliseconds. Default 0.
excludeNoA CSS selector. Matching elements, and everything inside them, are left out of every rule.
insecureNoIgnore TLS certificate errors, for a dev server with a self-signed certificate. Default false.
maxNodesNoElements shown per rule at detail "elements". Default 5. 0 shows every element.
viewportNoViewport as WxH. Default 1440x900. Responsive pages serve different content per width, so the width is part of the result.
timeoutMsNoNavigation timeout in milliseconds. Default 30000.
bestPracticesNoAlso run the best-practice rules (heading order, landmarks, redundant ARIA). Default false.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observedv0.1.0

TDQS

A4.3/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations exist, so the description carries the full burden and does well: it discloses the load-then-run mechanism against the live DOM, the result shape (severity, success criteria, CSS paths, fix, unjudged-rule count), and that results vary with viewport width and scroll state "by design." Remaining gaps are auth/environment prerequisites (e.g. that headless Chrome must be available) and whether page loads have side effects.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with purpose and standard scope, then mechanism, then caveats; no sentence is filler. It is dense and long-ish for a description, but the density is informational rather than redundant.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 13-parameter, annotation-free tool with no output schema, the description substitutes well: it explains the execution model, the meaning of key result fields, and the state dependence of results. Only environment prerequisites and the absence of a conformance-claim caveat beyond the wcag3 note keep it from being fully complete.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3, but the description adds genuine cross-parameter meaning: viewport width changes responsive content and scrolling pulls in lazy content, so both are part of the result and must be reported. That framing makes the viewport/scroll parameters more than their schema text alone conveys.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ("Audit a URL against WCAG 2.2 A and AA") plus the engine and environment ("pour engine in headless Chrome"). It is clearly distinguishable from pour_focus_order and pour_screenshot, which do not audit rules, and it names the alternative standard mode (wcag3).

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives concrete usage context: works on a dev server or any URL, wcag3 should be "left unset unless WCAG 3 is asked for," and it warns the caller to report which viewport/scroll state was used. It does not explicitly route between this tool and its siblings, so it stops short of full when/when-not/alternatives guidance.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server

Other Tools