Security audit

APEX IA Scanner

Security checks across malware telemetry and agentic risk

Overview

This package is advertised as a Binance Futures scanner, but it also includes code that can automatically place leveraged futures trades using embedded Binance credentials.

Treat this as a high-risk trading bot package, not just a scanner. Do not run it with real exchange permissions unless the trading files are removed or audited, hardcoded keys are revoked and replaced with least-privilege user-managed credentials, auto-trading is disabled by default, and every account-changing action requires clear user approval.

SkillSpector

By NVIDIA

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 (61)

Description-Behavior Mismatch

High

Confidence: 99% confidence
Finding: The skill is presented as a scanner, but it contains authenticated trading logic that sets leverage, opens futures positions, places take-profit orders, and closes positions. This mismatch is dangerous because users or platforms may grant it lower scrutiny than a trading bot, while it can still execute leveraged orders and materially affect exchange accounts.

Context-Inappropriate Capability

Critical

Confidence: 100% confidence
Finding: The file embeds full Binance API credentials directly in source code and uses them for signed account and order requests. Hardcoded secrets can be extracted by anyone with code access, enabling unauthorized account access, balance retrieval, leverage changes, and order placement; in a trading context this is especially severe.

Context-Inappropriate Capability

High

Confidence: 98% confidence
Finding: Auto-trading is enabled by default, so detected signals can trigger futures market orders without an explicit activation step by the user. In a skill advertised as a scanner, this creates unexpected account-changing behavior and can lead to unwanted leveraged trades, losses, or liquidation risk.

Intent-Code Divergence

Medium

Confidence: 93% confidence
Finding: The comments and UI emphasize scanning while the actual behavior performs aggressive leveraged trading with no stop loss by default. This misleading presentation lowers user vigilance and can cause operators to underestimate the real financial and operational risk of running the skill.

Description-Behavior Mismatch

High

Confidence: 98% confidence
Finding: The manifest presents this skill as a scanner, but the code contains full order-placement and position-management logic for Binance Futures. This mismatch is dangerous because users or higher-level agents may grant the skill lower-risk permissions while it is actually capable of executing leveraged trades and changing account state.

Context-Inappropriate Capability

High

Confidence: 97% confidence
Finding: A scanner-only skill should not need signed account access, balance retrieval, or authenticated Binance API requests. Including these capabilities expands the trust boundary and enables access to sensitive account data and trade-capable endpoints, making accidental or unauthorized account actions possible.

Context-Inappropriate Capability

High

Confidence: 96% confidence
Finding: The interactive key handler can switch the program into automatic trading mode, which exceeds the declared scanner role. In an agent or terminal environment, a hidden or poorly understood mode toggle can cause the tool to place trades unexpectedly once market signals are detected.

Description-Behavior Mismatch

High

Confidence: 98% confidence
Finding: The skill is presented as a scanner, but its configuration enables automated futures trading with leverage and account-affecting behavior. This mismatch is dangerous because users may grant trust or run the tool expecting passive analysis, while it can actively place risky market orders and modify trading settings.

Description-Behavior Mismatch

High

Confidence: 99% confidence
Finding: This code performs authenticated Binance futures actions including leverage changes and order placement, which goes far beyond scanning. In the context of a purported scanner skill, these side effects can directly alter a user's exchange account and create financial loss, especially with leveraged derivatives.

Description-Behavior Mismatch

High

Confidence: 99% confidence
Finding: The skill automatically converts detected signals into trades when autoTrade is enabled, and autoTrade is enabled by default. This is dangerous because trading decisions are executed without per-trade confirmation, allowing rapid unintended leveraged exposure from a tool advertised as a scanner.

Context-Inappropriate Capability

High

Confidence: 100% confidence
Finding: The file hardcodes a Binance API key and secret directly in source code, exposing sensitive credentials to anyone with access to the skill. Even if intended for testnet, embedded secrets are unsafe, easily reused incorrectly, and unjustified for a scanner-only tool.

Context-Inappropriate Capability

Medium

Confidence: 87% confidence
Finding: The skill persistently writes operation logs into the user's home directory even though it is described as a scanner. Persistent logging can expose trading activity, account-related errors, and sensitive operational details on disk without clear user consent or retention controls.

Intent-Code Divergence

Medium

Confidence: 90% confidence
Finding: The comment and surrounding labeling frame the module as an SMA scanner, while the implementation nearby participates in automated leveraged trade execution. This deceptive framing increases the risk that reviewers or users misunderstand the true behavior and approve or run hazardous code.

Description-Behavior Mismatch

High

Confidence: 98% confidence
Finding: The skill is ներկայացված as a market scanner, but this code places and closes Binance Futures orders, materially expanding its capabilities from passive analysis to active trading. That mismatch is dangerous because users may grant trust or run the skill expecting read-only behavior while it can initiate leveraged trades and incur financial loss.

Description-Behavior Mismatch

High

Confidence: 99% confidence
Finding: The code enables an automatic trading mode that executes trades from generated signals, even though the skill is described only as a scanner. In this context, hidden execution capability is especially risky because a user may trigger automated financial actions without understanding that the tool can place futures orders.

Context-Inappropriate Capability

High

Confidence: 100% confidence
Finding: Hardcoded Binance API key and secret are embedded directly in the source, exposing sensitive credentials to anyone with file access. Because this skill also performs signed futures trading requests, credential compromise could allow unauthorized account access, order placement, and financial loss.

Description-Behavior Mismatch

High

Confidence: 99% confidence
Finding: The skill is described as a scanner, but this code places real Binance Futures orders, sets leverage, and manages positions. That mismatch is dangerous because users may run it expecting passive market analysis while it performs autonomous trading actions with financial risk.

Description-Behavior Mismatch

High

Confidence: 99% confidence
Finding: The main loop automatically executes trades for detected signals instead of only surfacing scan results. In the context of a purported scanner, this creates hidden behavior that can trigger repeated futures orders without informed user consent.

Description-Behavior Mismatch

High

Confidence: 98% confidence
Finding: The file behavior materially differs from the skill metadata: the manifest advertises an SMA 8/21 crossover scanner, but the code implements a different SMC strategy using PDH/PDL, structure-break, and FVG logic. This is dangerous because users may grant trust, run, or operationally depend on the skill under false assumptions, which is a security-relevant integrity and transparency failure for agent skills.

Description-Behavior Mismatch

High

Confidence: 99% confidence
Finding: The manifest advertises a Binance Futures scanner for SMA 8/21 crossovers, but the code configures and runs a different SMC strategy with automated trading enabled. This mismatch is dangerous because users and host systems may grant permission for passive market scanning while the skill actually performs leveraged order execution, creating undisclosed financial risk.

Description-Behavior Mismatch

High

Confidence: 99% confidence
Finding: This code places market, stop-loss, and take-profit orders and tracks live positions, which materially exceeds a scanner-only description. Hidden execution capability is dangerous because it can trigger real or testnet leveraged trades without informed consent, and the same code path could be switched from demo to live trading by changing one flag.

Context-Inappropriate Capability

Critical

Confidence: 100% confidence
Finding: Hardcoded Binance API credentials embedded in source code expose authenticated trading capability to anyone who can read, copy, or reuse the file. In a trading context this is especially dangerous because exposed keys can be abused for unauthorized account actions, and the skill's stated purpose does not justify bundling exchange credentials at all.

Context-Inappropriate Capability

Medium

Confidence: 82% confidence
Finding: The skill writes persistent logs to the user's home directory even though the manifest only describes scanning. While not as severe as credential exposure, this creates undisclosed local data retention and can leak trading activity, balances, symbols, or operational details to other local users or backup systems.

Intent-Code Divergence

High

Confidence: 98% confidence
Finding: The file labels the strategy as SMC using PDH/PDL, Choque, and FVG, directly contradicting the manifest's stated SMA 8/21 crossover scanner purpose. This inconsistency increases the risk of deceptive deployment because reviewers or users may approve one trading logic while the code runs another, undermining informed consent and safety review.

Description-Behavior Mismatch

High

Confidence: 98% confidence
Finding: The skill metadata claims it is a scanner, but this code opens and closes Binance Futures positions, which is materially different behavior. This mismatch is dangerous because users may grant trust, permissions, or credentials expecting passive analysis while the code can actively trade and create financial loss.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.dangerous_exec, suspicious.exposed_secret_literal

Shell command execution detected (child_process).

Critical

Code: suspicious.dangerous_exec
Location: realtime-scanner.js:155

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: apex-ia-aggressive.mjs:13

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: apex-ia-all.mjs:13

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: apex-ia-final-20x.mjs:53

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: apex-ia-final.mjs:13

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: apex-ia-robot.mjs:13

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: apex-ia-smc.mjs:53

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: apex-ia-trader.mjs:14