ClawHub
Back to skill

Security audit

longbridge-trader

Security checks across malware telemetry and agentic risk

Overview

This appears to be a legitimate Longbridge trading helper, but it gives an agent live brokerage-account authority with incomplete safeguards and disclosure.

Install only if you intend to let the agent interact with a Longbridge/Longport brokerage account. Prefer quote-only or read-only credentials where possible, avoid trade permissions unless needed, and require explicit confirmation for every order placement, modification, or cancellation with symbol, side, quantity, price, and account shown first.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Vague Triggers

High
Confidence
96% confidence
Finding
The skill explicitly requires triggering on extremely broad securities-related mentions, including casual questions like stock prices or account balance. In an agentic environment, overly broad mandatory triggering can cause the trading skill to intercept ambiguous user requests and route them into a capability set that includes account access and order placement, increasing the risk of unintended financial actions or privacy exposure.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
This documentation exposes order-placement, cancellation, and modification APIs for live securities trading without any explicit warning, confirmation requirements, or guidance to distinguish simulation from real trading. In the context of an agent skill that must trigger on casual trading-related queries, this increases the chance that an LLM-based agent could invoke account-impacting actions from ambiguous or insufficiently confirmed user input, leading to unintended real orders and financial loss.

Missing User Warnings

Low
Confidence
82% confidence
Finding
The file documents access to sensitive financial data such as account balances, positions, executions, and cash flow without any privacy warning, data-minimization guidance, or mention of authorization boundaries. In a trading assistant skill that is expected to answer casual questions about account state, this makes inadvertent disclosure more likely if the agent surfaces more account data than necessary or responds in an insecure context.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.