Browser Agent (yubao2000) (yubao2000) MCP Server
@yubao2000/mcp-browser-agent
npm
v0.5.0
Published by @yubao2000 — no publish provenance and no vendor-owned scope, so the publisher could not be verified. The repository link below is self-declared.
Lets AI agents operate a browser like a human, with around 30 tools including cookie persistence, PDF export, and tab management.
Every server starts at 100. These are the exact deductions the deterministic engine applied — each one reproducible, none of it an opinion or an LLM's guess:
| Points | What was found | Category |
|---|---|---|
| −27 | Dynamic evaluation of a non-literal value ×3 MTC-SRC-010 | injection |
Capability observations and info notes are shown under Findings but never scored.
Open any row's finding below for the file, line and evidence behind the deduction.
Findings 8
high Dynamic evaluation of a non-literal value (dist/index.js)MTC-SRC-010
In the server's implementation (`dist/index.js:654`): Evaluating a runtime value as code (rather than a fixed literal) executes whatever reaches it — a direct RCE primitive, and almost never necessary in legitimate code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: eval(
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server dist/index.js
high Dynamic evaluation of a non-literal value (dist/pro/features.js)MTC-SRC-010
In the server's implementation (`dist/pro/features.js:94`): Evaluating a runtime value as code (rather than a fixed literal) executes whatever reaches it — a direct RCE primitive, and almost never necessary in legitimate code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: eval(
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server dist/pro/features.js
high Dynamic evaluation of a non-literal value (src/pro/features.ts)MTC-SRC-010
In the server's implementation (`src/pro/features.ts:135`): Evaluating a runtime value as code (rather than a fixed literal) executes whatever reaches it — a direct RCE primitive, and almost never necessary in legitimate code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: eval(
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server src/pro/features.ts
high Dynamic code execution in server code (dist/index.js)MTC-SRC-001
In the server's implementation (`dist/index.js:654`): Evaluating strings as code is the most direct RCE primitive; if any tool input reaches it, the server executes attacker-chosen code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: eval(
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server dist/index.js
high Dynamic code execution in server code (dist/pro/features.js)MTC-SRC-001
In the server's implementation (`dist/pro/features.js:94`): Evaluating strings as code is the most direct RCE primitive; if any tool input reaches it, the server executes attacker-chosen code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: eval(
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server dist/pro/features.js
high Dynamic code execution in server code (src/pro/features.ts)MTC-SRC-001
In the server's implementation (`src/pro/features.ts:135`): Evaluating strings as code is the most direct RCE primitive; if any tool input reaches it, the server executes attacker-chosen code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: eval(
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server src/pro/features.ts
medium Hardcoded egress to an external endpoint (dist/pro/license.js)MTC-SRC-003
In the server's implementation (`dist/pro/license.js:44`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: fetch("https://api.lemonsqueezy.com/v1/licenses/validate"
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server dist/pro/license.js
medium Hardcoded egress to an external endpoint (src/pro/license.ts)MTC-SRC-003
In the server's implementation (`src/pro/license.ts:67`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.
Evidence: fetch("https://api.lemonsqueezy.com/v1/licenses/validate"
Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.
Location: server src/pro/license.ts
What this scan could not see
- No tools were enumerated, so prompt-injection, capability and toxic-flow analysis had no tool surface to inspect. To grade a package’s real runtime tools, scan the running server: --command "npx -y <package>".
Versions
Scan history per published version. The engine is deterministic — the same version always yields the same score, so a changed score means the package itself changed.
| Version | Score | Findings | Engine | Scanned |
|---|---|---|---|---|
v0.5.0 latest |
C 73/100 | 8 | 1.5.0 | 2026-07-22 |
Embed this score
Show this server's live Trust Score in your README, docs or website. The badge is served straight from the registry and updates automatically after every rescan — no API key needed. It links back to this page, so anyone who sees the grade can also read the findings behind it instead of taking a number on faith.
Verify this score yourself
The score above is reproducible: the same package version always yields the same result. Run it locally or over the free API — no account, no LLM, fully deterministic.
npx mcptrustchecker scan @yubao2000/mcp-browser-agent --onlineOther implementations of Browser Agent (yubao2000) 1
Independent packages implementing the same tool, scanned with the same engine. Compare all 2 side by side →
More in Browser Automation
Security scan results for the 3dstreet MCP server.
Security scan results for the Accessibility Scanner MCP server.
Security scan results for the Adspower Browser MCP server.
Security scan results for the Agent Browser MCP server.
Security scan results for the Agent Browser MCP server.
Security scan results for the Agent Browser Mcp Codex MCP server.