@agentdeskai/browser-tools-mcp

npm / sealed 2026-06-17 / No. 960c83dc

Does @agentdeskai/browser-tools-mcp MCP work in Claude Code and Cursor? Throne executed @agentdeskai/browser-tools-mcp mcp in a single-use Firecracker microVM and replayed both recorded client behaviours. Compatibility verdict: fit. Is @agentdeskai/browser-tools-mcp mcp safe? Static security scan: 1 finding(s) to review under security ruleset v1. The full per-step results, security findings, scan date, and evidence hash are below.

re-scan this server gate it in your CI

> throne registry @agentdeskai/browser-tools-mcp sealed

receiptsealed evidence

scan id: 960c83dc28ba4e9bb4258c059c4a60df
target: @agentdeskai/browser-tools-mcp
sealed at: 2026-06-17 11:25:24Z
evidence hash: sha256:298dc73f889c37c10450d0fd625221039a78f018d43b285d0a2f47d415bbb72b

01connectPASS905ms

initialize ok: server Browser Tools MCP 1.2.0, negotiated protocolVersion 2025-11-25, capabilities ['tools']

02discoverPASS7ms

supported: tools/list (14 tools); method not found (tolerated): resources/list, prompts/list

03validate_schemasPASS6ms

all 14 tool inputSchemas are valid JSON Schema

04smoke_test_toolsPASS189ms

called 10 of 10 tools (cap 10, strategy=serial): 0 ok, 10 returned tool-level errors (expected for synthesized args)

05error_handlingPASS5.0s

structured error responses, connection survived, wrong_type_args: getConsoleLogs:accepted (returned result); getConsoleErrors:accepted (returned result); getNetworkErrors:accepted (returned result); getNetworkLogs:accepted (returned result); takeScreenshot:accepted (returned result); unknown_method: error(-32601); invalid_id: no response (silent), connection alive; oversized_input: accepted (returned result); recovery: accepted (returned result); cancellation: accepted (returned result); old_version: accepted old version 2024-10-07, negotiated 2024-10-07, server may have stale protocol handling

06streamingNOT_APPLICABLE1ms

no streaming-capable tool declared by this server

07resource_lifecycleNOT_APPLICABLE3ms

resources capability not declared

08concurrent_callsWARN36ms

ladder 1 overlapping getConsoleLogs calls: max_observed_stable_concurrency=1; all ids answered exactly once, but every call returned an error (tool-level isError result), the tool may need runtime dependencies this sandbox does not provide

09reconnectPASS848ms

transport closed and relaunched; re-handshake ok (protocolVersion 2025-11-25, first session was 2025-11-25)

01connectPASS1.7s

initialize ok: server Browser Tools MCP 1.2.0, negotiated protocolVersion 2025-11-25, capabilities ['tools'] [request_strategy=pipelined, connection_count=2, second connection established]

02discoverPASS8ms

supported: tools/list (14 tools); method not found (tolerated): resources/list, prompts/list; notes: connection 2 confirms 14 tools

03validate_schemasPASS3ms

all 14 tool inputSchemas are valid JSON Schema

04smoke_test_toolsPASS147ms

called 10 of 10 tools (cap 10, strategy=pipelined): 0 ok, 10 returned tool-level errors (expected for synthesized args)

05error_handlingPASS4.8s

06streamingNOT_APPLICABLE0ms

no streaming-capable tool declared by this server

07resource_lifecycleNOT_APPLICABLE1ms

resources capability not declared

08concurrent_callsWARN224ms

ladder 1/2/4/6 overlapping getConsoleLogs calls: max_observed_stable_concurrency=6; all ids answered exactly once, but every call returned an error (tool-level isError result), the tool may need runtime dependencies this sandbox does not provide

09reconnectPASS1.6s

transport closed and relaunched; re-handshake ok (protocolVersion 2025-11-25, first session was 2025-11-25)

chatgpt desktop calibratingemulation profile pending real-traffic captureCOMING SOON

SECURITY: REVIEW / 1 finding(s) / review recommended

LOWTHR-NET-05 / Hardcoded outbound endpoints

6 non-local endpoint host(s) referenced in code, verify each is expected for this server's purpose: dminhvu.com (package/dist/mcp-server.js:562), ik.imagekit.io (package/dist/mcp-server.js:565), nextjs.org (package/dist/mcp-server.js:846), www.google.com (package/dist/mcp-server.js:864), www.linkedin.com (package/dist/mcp-server.js:581), www.sitemaps.org (package/dist/mcp-server.js:864)

package/dist/mcp-server.js:562

VERDICT: FIT14 of 18 steps exercised across 2 clientsSANDBOXED RUN, submitted server executed in a disposable microVM, compatibility: 0 fail / 2 warn (14 of 18 steps exercised across 2 clients) / security: review, 1 finding(s), 0 highsealed by THRONE / No. 960c83dc / 2026-06-17

scopeattestation tuple

target: @agentdeskai/browser-tools-mcp npm
engine: sandboxed
claude code: calibration: partial recorded=9,spec=0,assumed=9
cursor: calibration: partial recorded=11,spec=0,assumed=7
chatgpt desktop: calibration: unavailable recorded=0,spec=2,assumed=16
steps exercised: 14 of 18
test suite: v1.0.0
security ruleset: v1.0.0
sealed at: 2026-06-17 11:25:24Z
evidence hash: sha256:298dc73f889c37c10450d0fd625221039a78f018d43b285d0a2f47d415bbb72b
valid until: 2026-09-15

wear the crown

paste this in your README. it renders the live verdict and links back to this record. if a release ever breaks the verdict, the badge says so on its own.

[![THRONE: FIT](https://api.usethrone.dev/api/badge/%40agentdeskai%2Fbrowser-tools-mcp)](https://usethrone.dev/server/agentdeskai-browser-tools-mcp)

executed in a disposable microVM, created for this scan and destroyed after it. nothing outlives a run.

maintainer of this server? challenge this record: hello@usethrone.dev. tell us what we got wrong and we re-run it in the open.

this page renders the stored record of a real run. nothing on it is asserted without the execution that proved it.