Verified Trading Bot Reputation: Building Cryptographic PnL Proof

• Outcome: Review
• Latest audit: May 29, 2026 at 9:33 PM UTC
• Version:
  1.3.1

Overview

The skill is a non-executable guide, but it under-discloses that examples use live GreenHelix APIs, bearer tokens, a sensitive signing key, and trading-performance data submissions.

Install only if you understand this is an API-integration guide, not a sandbox-only tutorial. Use a dedicated test signing key first, do not reuse a wallet or exchange key, verify whether each snippet points to sandbox or production before running it, and assume submitted trading metrics may become discoverable or retained by GreenHelix.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Description-Behavior Mismatch

Low

• Confidence: 96% confidence
• Finding: The guide claims it is non-executable educational content, yet it embeds complete curl and Python examples that perform real network operations. This mismatch can mislead users and downstream systems into underestimating the operational and data-sharing risk of using the skill.

Intent-Code Divergence

Medium

• Confidence: 97% confidence
• Finding: The documentation states that all examples use the sandbox and require no API key, but many examples target the production API and include Bearer-token authentication. This inconsistency can cause users to send real credentials and real trading data to production endpoints under the false assumption they are in a harmless test environment.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The examples submit agent identifiers, trading performance metrics, timestamps, and signatures to an external service without nearby disclosure of privacy, retention, or third-party sharing implications. In a trading context, this data can reveal strategy behavior, performance history, and business-sensitive metadata.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill references an AGENT_SIGNING_KEY and demonstrates loading raw private key material, but it does not prominently warn that this credential is highly sensitive and identity-bearing. Mishandling the key could let an attacker forge submissions, impersonate the bot, poison reputation records, or undermine any claimed cryptographic trust model.

VirusTotal

65/65 vendors flagged this skill as clean.