ClawHub
Back to skill

Security audit

Go Trader

Security checks for vulnerabilities and agentic risk

Overview

This skill is meant to control a crypto trading bot, but it exposes high-impact trading/service actions and raw logs with weak safeguards and some documentation mismatches.

Install only if you intentionally want an agent to inspect and control a local go-trader service. Before using it with live funds, add explicit confirmations for stop/restart/live/reset actions, clarify what emergency stop actually does to open positions, restrict log access, and verify that only authorized users can run the service-control script.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (5)

External Script Fetching

High
Category
Supply Chain
Content
case "$COMMAND" in
    status)
        echo "📊 Trading Status:"
        curl -s localhost:8099/status 2>/dev/null | python3 -m json.tool || echo "❌ go-trader not responding"
        ;;
    health)
        echo "🏥 System Health:"
Confidence
90% confidence
Finding
Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill exposes destructive trading controls like "Emergency stop all" and "Reset trading state" without clearly describing consequences, safeguards, or approval requirements. In a live trading environment, accidental or socially engineered invocation could halt strategies, close positions, or erase state needed for safe recovery, causing financial loss and operational disruption.

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The top-of-file comments describe this as a connection layer to Telegram via OpenClaw. However, the implementation contains no Telegram or OpenClaw integration logic; it only dispatches local commands to curl localhost, journalctl, and systemctl. This is an active documentation-to-code mismatch rather than a mere omission.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The script exposes disruptive service-management actions like stop, start, and restart with no confirmation, authorization guard, or safety interlock. In an agent/automation context, this increases the risk of accidental or unauthorized invocation causing denial of service for the trading system at sensitive times.

Vague Triggers

Low
Confidence
90% confidence
Finding
The trigger "Show trading logs" is overly broad and lacks scope restrictions, which can cause the agent to reveal sensitive operational data in response to generic troubleshooting requests. In a trading bot context, logs may contain API errors, account identifiers, strategy behavior, or other sensitive telemetry, so ambiguous invocation increases the risk of unintended disclosure.

Static analysis

No suspicious patterns detected.