ClawHub
Back to skill

Security audit

Minimax Monitor

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent MiniMax quota dashboard, but it automatically reads local API credentials and calls MiniMax on page load despite saying this only happens after user confirmation.

Review before installing. This skill is intended to monitor MiniMax usage, but opening the dashboard may immediately use ~/.mmx/config.json and contact MiniMax rather than waiting for the credential-load confirmation. Only run it if you are comfortable with automatic local API-key use, outbound quota polling, and a localhost service on port 9877.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
The file imports execSync and exec from child_process even though no shell execution is used anywhere in the implementation. Unused command-execution capability expands the attack surface and normalizes dangerous permissions for a monitoring service, increasing the risk of later abuse or accidental introduction of command injection paths.

Intent-Code Divergence

Medium
Confidence
90% confidence
Finding
The startup path immediately calls startMon(), which triggers fetchQuota() on page load despite comments stating that config/credentials should not be read before explicit user action. Because the backend is documented to read ~/.mmx/config.json on demand when servicing /api/token_plan, this can cause implicit local credential use and outbound quota/API activity without informed user consent.

Static analysis

No suspicious patterns detected.