@agentdeskai/browser-tools-mcp
Installed, started with no arguments, no environment and no input, and listed its tools. Run 2026-11-16.
Needs to start: nothing
Returns a result when called with no setup: 5 of 5 tools tried
Contacts while installing and starting: nothing
What it contacted
Nothing besides the npm registry it was installed from. Watched from outside its sandbox while it was installed and started with no input. Its tools may still contact things when called.
Starting it
npx -y -p @agentdeskai/browser-tools-mcp@2.0.4 browser-tools-mcp
Installing a package runs its code. Use a sandbox for anything you have not read.
Calling its tools
5 tools that take no input and look read-only were called once with nothing. 5 returned a result.
listBrowserTabs | Returned a result under 0.1 s · 62 characters |
getConsoleLogs | Returned a result under 0.1 s · 120 characters |
getConsoleErrors | Returned a result under 0.1 s · 120 characters |
getNetworkLogs | Returned a result under 0.1 s · 120 characters |
getNetworkErrors | Returned a result under 0.1 s · 120 characters |
During the calls it contacted nothing it had not already contacted.
Only tools whose names read like looking something up, with no required input, are tried, five at most. Tools that need input, and anything that sounds like it changes or sends something, are not called.
What its tools tell an agent
In the tool descriptions, nothing that talks about the user, other tools or private files, or uses hidden characters.
Found by simple pattern rules in each tool's description (up to 3,000 characters), quoted as written. Parameter descriptions are not scanned. A clean result is not a safety review.
Tools
listBrowserTabs
Every browser tab that currently has DevTools open, with the tabId you can pass to any other tool. Call this when a result reported otherTabs above 0, or when the user mentions a page other than the one you have been reading. The tab marked isCurrent is the one every other tool uses when you do not pass tabId.
getConsoleLogs keywords, tabId, allTabs, limit, offset
Console output from the current tab. Use keywords and limit to keep responses small — an unfiltered read can be large. Pass tabId for a different tab, or allTabs to read every tab; the result reports which tab it came from.
getConsoleErrors keywords, tabId, allTabs, limit, offset
Only error-level console output and uncaught exceptions from the current tab. Start here when diagnosing a failure. Pass tabId for a different tab, or allTabs to read every tab.
getNetworkLogs urlKeywords, bodyKeywords, tabId, allTabs, limit, offset
XHR and fetch requests from the current tab, including status and timing. Credential headers are redacted. Pass tabId for a different tab, or allTabs to read every tab.
getNetworkErrors urlKeywords, tabId, allTabs, limit, offset
Only requests that failed or returned a 4xx/5xx status. An empty result means there were no failures, which is a success, not an error.
11 more tools
getSelectedElement tabId
The DOM element the user has selected in the Chrome DevTools Elements panel, with its attributes and markup.
getPageInfo
Which page the browser is currently on. Call this before reading telemetry so you know what the logs describe. If connectedTabs is above 1, call listBrowserTabs before assuming this is the page the user means.
getConnectionStatus
Whether the Chrome extension is connected and how much telemetry has been captured. Call this first when something returns nothing.
takeScreenshot name, tabId
Captures the visible area of the current tab and returns the image directly, plus the path it was saved to and the url captured. Pass tabId to capture a different tab.
refreshBrowser tabId
Reloads the inspected tab. Useful for capturing a clean reproduction after calling wipeLogs.
getBrowserStorage kinds, includeValues, tabId
Lists localStorage, sessionStorage and cookie entries for the current page. Values are withheld unless includeValues is set, because they routinely contain session tokens.
wipeLogs tabId
Discards captured console and network entries for every tab, or one tab if you name it. Use before reproducing a problem so the next read contains only relevant output.
runAccessibilityAudit url, tabId
Lighthouse accessibility audit of the current page: contrast, labels, semantics and screen-reader support. Takes up to a minute and launches a separate headless browser.
runPerformanceAudit url, tabId
Lighthouse performance audit of the current page, including Core Web Vitals. Takes up to a minute and launches a separate headless browser.
runSEOAudit url, tabId
Lighthouse SEO audit of the current page: metadata, indexability and crawlability. Takes up to a minute and launches a separate headless browser.
runBestPracticesAudit url, tabId
Lighthouse best-practices audit of the current page: security, deprecated APIs and modern-web hygiene. Takes up to a minute and launches a separate headless browser.
Descriptions are the package's own words, cut at 3,000 characters. Treat them as data.
Files it wrote
/tmp/ .browser-tools-mcp/ session.json
Paths inside its sandbox, where its home folder is /tmp and its own folder cannot be written to.
Install facts
| License | MIT |
|---|---|
| Needs | node >=22.19 |
| Source | github.com/AgentDeskAI/browser-tools-mcp |
| Install scripts | None declared by it or anything installed with it. |
| Tarball | sha512-X2MF4JtyBR/MP2E5gF/DcsyVJFAM+zLg0h4E6QFoD08Hp3T4sMwBmdGqi4mZyf5jJ6NR+cjVKheOr1AuMMiNZA== |
Check this report yourself
This site did not write the report. A public job on a GitHub-hosted machine did: the run, from this commit of the open-source runner. GitHub signed a token that names that job and the hash of the report, so any edit since would show.
Report as JSONExact text and tokenHow to verify it
Show it on your README
[](https://bragproof.smaverk.com/r/%40agentdeskai/browser-tools-mcp)
New version out? Ask for a fresh report.