MCP Registry

@iflow-mcp/mobilenext-mobile-mcp

a20a2d86-a0ed-5e04-8860-c4b83541fd01 | Store v0.0.1
77/ 100
HIGH risk
AI review verdict
Review before use

The AI review rates the findings as likely false positive, but the risk score (77/100) still counts them.

Analysis record

Analysed
2 months ago 26 August 2026
Version
v0.0.1
Artifact
SHA256 ad770f69…bcc30b
Detail
Full list

Evidence ledger

Ranked by severity · findings with a source location link to the code viewer

23 findings

Finding Categories

1
Secrets
21
Network

MCP Server Analysis

MCP servers expose tools and resources to AI assistants. Unlike browser extensions, they run as standalone processes with direct system access. Tool definitions are analyzed for prompt injection, data exfiltration, and tool poisoning patterns.

AI Security Report

AI Security Review

Evidence context: threat category none; evidence quality moderate.

Security Analysis: @iflow-mcp/mobilenext-mobile-mcp

Tool Poisoning Assessment

No tool-poisoning findings detected. The findings summary explicitly shows "tool-poisoning":"0". This is the defining threat for MCP servers, and there is no evidence of hidden AI manipulation directives, invisible Unicode characters, or XML-style instruction tags in tool descriptions. The package does not exhibit the tool-poisoning patterns that characterize malicious MCP servers.

Credential and Network Access

Network activity is legitimate and expected. All 21 network findings are NET-FETCH calls in lib/webdriver-agent.js at lines 15, 26, 107, 122, 148, 291, 347, 381, and 393. This file is a webdriver agent, and fetch calls are the expected mechanism for mobile automation tools to communicate with WebDriver servers. There is no evidence of exfiltration architecture.

No suspicious credential access. The findings show only 1 secret finding with no details indicating sensitive path targeting. There are no credential-access findings targeting .ssh/, .aws/credentials, .kube/config, or specific secret names like GITHUB_TOKEN or AWS_SECRET_ACCESS_KEY. The absence of the credential-harvest-plus-network-call combination rules out exfiltration.

IOC Analysis

All detected URLs are benign infrastructure. The three XIOC-URL findings reference:

  • https://github.com/orgs/mobile-next/projects/3 - GitHub project page
  • https://nodejs.org/en/download/ - Node.js download page
  • https://github.com/mobile-next/ - GitHub organization page

These are legitimate development infrastructure URLs, not exfiltration endpoints. The 94 total IOC findings are consistent with the known false-positive patterns from the IoC extractor matching property access chains and benign URLs in bundled code.

Strongest Counterargument

The package has 167 total findings with 94 IoC matches, which might suggest malicious activity. However, the MCP threat model explicitly states that IoC count alone is meaningless in MCP packages due to bundled dependencies. The critical indicators are all absent: zero tool-poisoning findings, zero malware signatures, zero credential theft patterns, and zero calls to suspicious domains. The network activity aligns with the stated purpose (webdriver agent), and the URLs are documented infrastructure. High finding counts from code-smell rules and IoC extraction noise do not constitute evidence of malice.

Conclusion

This is a legitimate mobile automation MCP server. The findings represent expected infrastructure references and code-smell noise, not malicious behavior.

Key Reasons

  • Zero tool-poisoning findings - the defining MCP threat is absent
  • Zero malware-signature findings
  • Network calls are in webdriver-agent.js, expected for mobile automation
  • All IoC URLs are benign (GitHub, Node.js)
  • No credential-access findings targeting sensitive paths

False Positive Considerations

  • IoC extractor matching benign infrastructure URLs (GitHub, nodejs.org)
  • Code-smell findings from standard Node.js patterns (35 findings)
  • High finding count from bundled dependencies
  • NET-FETCH findings in legitimate webdriver agent code

Reviewed 2026-04-27 · version 0.0.1; recommended action: no action; model confidence 85%.

Source Code Viewer

Source Code Not Available

Source code is not available for this version of the extension.

About This Extension

Mobile MCP

Frequently Asked Questions