agent-browser MCP server
com.scriptmasterlabs/agent-browserscriptmasterlabs.comstreamable-httpv2.0.3Scanned
Chromium browser agents pay per action via x402 — USDC on Base, USDC on Solana, or XRP on XRPL.
6 tools: 2 can delete or cancel things, 2 change things and 2 only read.
- Reads private data · 0
- Reads text others write · 5
- Can send data out · 1
- No findings
Tools listed without sign-in. Remote posture
With Anlyon
2 tools here can delete or cancel things. Govern them.
- Anlyon lists the tools and adopts none.
- Each tool you adopt gets approvals and receipts.
- Your agent never holds the key.
A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.
What the agent-browser MCP server can do
The agent-browser MCP server is published by scriptmasterlabs.com, a domain the MCP Registry verified. It has 6 tools. It can delete or cancel things (2 tools: browser_act and browser_close) and create or change data (2 tools: browser_navigate and browser_extract). 2 tools only read.
5 tools bring in text other people can write (browser_navigate, browser_extract, browser_snapshot and 2 more); 1 tool can send data out of your system (browser_navigate).
None of the fixed checks matched anything in its tool definitions. That is not a safety guarantee.
Its remote endpoint lists its tools without sign-in, and its tools changed 2 times since Anlyon Radar first saw them.
Last scanned Nov 18, 2026, version 2.0.3.
Tools
Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.
6 tools
Deletes or cancels 2Claude Code: deny
browser_actUse for: interacting with the page — click, fill, type, select, submit, press.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it destructive (destructiveHint: true).
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Use for: interacting with the page — click, fill, type, select, submit, press. Returns the resulting page title, URL and text so you can verify the action landed. Don't use when: the target element is not on the current page — navigate or snapshot first. This tool can change remote state (it can submit forms).
Annotations the server declares
- readOnlyHint: false
- destructiveHint: true
- openWorldHint: true
browser_closeUse for: ending a session and freeing its browser.No findings
Why it is in this group
- The name "browser_close" says "close": a word for a change.
- The server marks it destructive (destructiveHint: true).
Description, as the server wrote it Use for: ending a session and freeing its browser. Sessions also expire automatically after 10 idle minutes. Free on every door — closing is never charged.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: true
- openWorldHint: false
Changes things 2Claude Code: ask
browser_extractUse for: pulling repeated elements (prices, titles, links) from the page a session is on, via a CSS selector, or the full page text when no selector is given.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Use for: pulling repeated elements (prices, titles, links) from the page a session is on, via a CSS selector, or the full page text when no selector is given. Returns one string per matched element. Don't use when: the page has not been loaded yet — call browser_navigate first.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- openWorldHint: true
Only reads 2Show 2 read-only tools
browser_screenshotUse for: visual verification — layout, images, charts, CAPTCHA presence, or evidence of page state.Reads text others writeNo findings
Why it is in this group
- The name "browser_screenshot" says "screenshot": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe name "browser_screenshot" says "screenshot" and "browser": it reads web pages, which anyone can write.
Description, as the server wrote it Use for: visual verification — layout, images, charts, CAPTCHA presence, or evidence of page state. Returns a PNG image (JPEG if the PNG exceeds 2.5MB). Don't use when: you only need the page's content or structure — browser_extract and browser_snapshot are far cheaper in tokens.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- openWorldHint: true
browser_snapshotUse for: understanding page structure cheaply and deterministically — returns the accessibility tree (roles, names, states) as compact YAML instead of pixels or raw HTML.Reads text others writeNo findings
Why it is in this group
- The name "browser_snapshot" says "snapshot": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe name "browser_snapshot" says "snapshot" and "browser": it reads web pages, which anyone can write.
- untrusted-inputThe description's first sentence says "html": it reads web pages, which anyone can write.
Description, as the server wrote it Use for: understanding page structure cheaply and deterministically — returns the accessibility tree (roles, names, states) as compact YAML instead of pixels or raw HTML. This is the lowest-token way to decide what to click or fill next. Don't use when: you need visual layout or images — use browser_screenshot.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- openWorldHint: true
This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.
Exfil surface
Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.
No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data.
- Reads private data0
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write5
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out1
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.
Remote posture
What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.
Looked at Nov 18, 2026, 16:23 UTC, unchanged since Oct 27, 2026. No credentials were sent and no tool was called. What is looked at.
- Reachable
- yes
- HTTP status of an anonymous initialize
- 200
- tools/list served without credentials
- yes
- Tools that change things, listed without credentials
- 4
- Protocol version
- 2025-06-18
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- not observed
- Authorization servers
- not observed
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- not observed
- Client ID metadata documents
- not observed
- HighThe whole server
The endpoint served tools/list without credentials, and the list has 4 tools that change things (2 destructive, 2 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://agent-browser-r7d2.onrender.com/mcp: tools/list returned 6 tools to an anonymous client, including browser_act (destructive), browser_close (destructive), browser_navigate (write), browser_extract (write)
Findings
Matches of fixed checks, with the text they matched.
No check matched anything in these tool definitions. That is not a safety guarantee.
History
Changes to its tool definitions and to its remote endpoint, newest first.
Tools5 changed.~5 changedView diff
- Changed:browser_actDeletes or cancelsdescription
- Changed:browser_extractChanges thingsdescription
- Changed:browser_navigateChanges thingsdescription
- Changed:browser_screenshotOnly readsdescription
- Changed:browser_snapshotOnly readsdescription
Tools6 changed.~6 changedView diff
- Changed:browser_actDeletes or cancelsdescription
- Changed:browser_closeDeletes or cancelsdescription
- Changed:browser_extractChanges thingsdescription
- Changed:browser_navigateChanges thingsdescription
- Changed:browser_screenshotOnly readsdescription
- Changed:browser_snapshotOnly readsdescription
Install
Pinned to version 2.0.3. Values in angle brackets are yours to fill in.
claude mcp add --transport http agent-browser https://agent-browser-r7d2.onrender.com/mcpclaude mcp add --transport http agent-browser https://agent-browser-r7d2.onrender.com/mcpRun in a terminal.
Claude Code permissions
Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.
{
"permissions": {
"allow": [
"mcp__agent-browser__browser_snapshot",
"mcp__agent-browser__browser_screenshot"
],
"ask": [
"mcp__agent-browser__browser_navigate",
"mcp__agent-browser__browser_extract"
],
"deny": [
"mcp__agent-browser__browser_act",
"mcp__agent-browser__browser_close"
]
}
}Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as agent-browser. Review it before you use it.
Capability tags behind 5 rules
- mcp__agent-browser__browser_navigate untrusted-input, outbound
- mcp__agent-browser__browser_extract untrusted-input
- mcp__agent-browser__browser_snapshot untrusted-input
- mcp__agent-browser__browser_act untrusted-input
- mcp__agent-browser__browser_screenshot untrusted-input
Pin
Keep the decision: fail CI when the tools change, follow every change, or show a badge.
Pin in CI
Fail a build when the tools you reviewed are not the tools the server serves today.
curl -s "https://radar.anlyon.com/api/v1/servers/com.scriptmasterlabs/agent-browser/verify?hash=5e58b55811114b4a289c93851071ebe5eb64a8fe2d1fd499ea8ed4d182649c2d" | jq -e '.changed == false'Pinned hash 5e58b5581111, from Nov 18, 2026, 20:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/com.scriptmasterlabs/agent-browser/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/com.scriptmasterlabs/agent-browser)Source: MCP Registry · repository. Synced Nov 5, 2026, 02:17 UTC.
agent-browser MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the agent-browser MCP server safe?
- No scan can promise that. What its definitions show: 6 tools, 4 of which can change, send or delete something, and 0 findings from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.
Can the agent-browser MCP server delete data?
- Yes. 2 tools can delete, cancel or revoke something: browser_act and browser_close.
Can the agent-browser MCP server do things that cannot be undone?
- No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.
What tools does the agent-browser MCP server have?
- 6 tools: 2 destructive, 2 write, 2 read. browser_act, browser_close, browser_navigate, browser_extract, browser_snapshot and browser_screenshot.
Does the agent-browser MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the agent-browser MCP server send data to outside services?
- 1 tool can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): browser_navigate.
Has the agent-browser MCP server changed its tools?
- Yes, 2 times since Anlyon Radar started tracking it, most recently on Nov 2, 2026. Every change is on this page and in its Atom feed.