ClawHub
Back to skill

Security audit

toasttab-mcp

Security checks for vulnerabilities and agentic risk

Overview

The skill provides disclosed, read-only Toast access, including personal account information when you are signed in.

Review the server and extension permissions before installing. For menu lookup alone, keep personal account tools disabled; enable them only when you want the agent to read your Toast account information.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

The skill extends restaurant and menu lookup into authenticated access to personal account data.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The manifest at L03 describes finding restaurants and reading menus through an order.toasttab.com tab. L21–L30 additionally instruct the agent to use a signed-in toast.app session to retrieve order history, full receipts, loyalty balances, and wallet balances, including prompting the user to sign in. These personal-account operations exceed the advertised restaurant-discovery and menu-reading scope.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.