Security audit

msx trade

Security checks across malware telemetry and agentic risk

Overview

This is a transparent MSX trading skill, but it gives an agent sensitive financial-data access and order-changing authority with some under-scoped safeguards.

Install only if you trust the publisher and the MSX API enough to let an agent use an API key with account and trading privileges. Prefer a scoped or read-only key where possible, confirm any order placement or cancellation yourself, and use MSX-specific prompts when requesting balances, history, or reports.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (6)

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The trigger phrases are broad enough to activate on common account- or finance-related language, which can cause this skill to load in conversations where the user did not intend to access trading or portfolio functionality. In a financial skill, unintended activation increases the chance of exposing or acting on sensitive account context, especially when paired with API-key-authenticated endpoints.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The file documents direct access to account profile, balance, and positions using only an API key, but provides no warning, consent flow, or guidance for handling sensitive financial data. In the context of a trading skill, this makes accidental or overbroad data access more dangerous because personal and portfolio information could be retrieved without clear user awareness or least-privilege constraints.

Vague Triggers

Medium

Confidence: 87% confidence
Finding: The trigger phrases are very broad and include generic terms like "history," "report," and "transactions," which can cause the skill to activate for unrelated requests. In a financial-trading context, mistaken routing is risky because it may surface or fetch sensitive trade and account-history data when the user did not clearly intend to access private financial records.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: This module exposes highly sensitive financial information, including deposits, withdrawals, balances, and executed trades, but provides no privacy warning, confirmation step, or minimization guidance. In a brokerage-like skill, that omission increases the chance of over-disclosure to the wrong user, in the wrong context, or with more detail than necessary.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The trigger list includes very broad terms like "buy", "sell", and "order", which can appear in non-trading conversations and cause the order-execution module to load inappropriately. In a financial trading skill, accidental activation increases the chance that the agent interprets unrelated user input as trading intent and progresses toward a consequential action.

Missing User Warnings

Medium

Confidence: 81% confidence
Finding: The cancellation section documents a destructive account action but lacks a requirement to confirm user intent before issuing DELETE /orders/{order_id}. Because cancelling an open or partially filled order can materially affect the user's position and execution strategy, omission of a confirmation safeguard creates risk of accidental or prompt-induced account changes.

VirusTotal

60/60 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.