How an AI agent uses Page audit and link checking
Accessibility, spelling and links — reported honestly, including what could not be checked.
Updated · T3rnel Browser 1.3.0
What the agent can do
- Flags missing alt text, heading-hierarchy skips, unlabelled form controls and ARIA misuse.
- Link checking distinguishes three outcomes rather than two: reachable, verifiably broken, and answered-but-unreadable.
- That third state is not a hedge. A cross-origin response is opaque to the browser —
statusreads as 0 whatever the server returned. Calling opaque "fine" marks every dead external link healthy; calling it "broken" marks every healthy one dead. Neither is a link check.
The tools
| Tool | Risk | Tier |
|---|---|---|
browser_audit | low | Free |
browser_check_links | low | Free |
browser_check_spelling | low | Free |
Calling it over MCP
{ "method": "tools/call", "params": { "name": "browser_audit", "arguments": {} } }Connect a client first: add the MCP Session Bridge. The extension then exposes 97 tools against your signed-in tabs.
Questions
Which tools does Page audit and link checking use?
browser_audit, browser_check_links, browser_check_spelling.
Is Page audit and link checking free?
Yes. Every tool listed here works on the free tier.
Does the agent have to ask before using Page audit and link checking?
Each tool has a risk level: browser_audit is low; browser_check_links is low; browser_check_spelling is low. The approval gate is off by default and can be armed in Settings.