ClawHub
Back to skill

Security audit

Motley — NYC neighborhoods, schools & admissions

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed read-only NYC information connector, with one caution around trusting its auto-approval advice for a remote MCP server.

Before installing, understand that your NYC-related queries may be sent to Motley's hosted MCP service. Do not rely solely on the skill's claim that tools are safe to auto-approve; keep normal MCP permission controls unless you independently trust the server. The artifact does not show account access, local file access, data mutation, or hidden execution.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
93% confidence
Finding

The skill explicitly instructs the agent that the remote MCP tools are 'safe to auto-approve' because they are 'read-only and idempotent.' That is a trust assertion about an external hosted service that the agent cannot independently verify at runtime, and auto-approval reduces human oversight for network calls and tool behavior changes over time. In this context, the risk is moderated because the described tools are informational and appear intended for read-only NYC data access, but the instruction still encourages unsafe delegation of approval decisions to the skill author.

Content

Scanner excerpt · SKILL.md (reported line 45)May include surrounding context.

text

Public and keyless — no API key, no account, nothing to install. Every tool is
read-only and idempotent, so it is safe to auto-approve. Tools return typed
`structuredContent` alongside their text: parse the object, don't re-parse JSON
out of prose.

Static analysis

No suspicious patterns detected.