ClawHub
Back to skill

Security audit

HFT Paper Trader Pro

Security checks for vulnerabilities and agentic risk

Overview

This is a paper-trading skill with disclosed local state files and no executable installer, but users should treat its automated trade examples as simulation-only and confirm write actions.

Install only if you want an agent to help run crypto paper-trading simulations. Before using broad prompts like trading all watchlist signals, confirm that it is not connected to real exchange credentials and decide where portfolio.json, journal.json, and observations.md should be written.

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
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Intent-Code Divergence

Medium
Confidence
89% confidence
Finding
The feature text says 'Correct SL placement: Math.max caps risk at 3%' and frames this as the implemented fix, but the dedicated stop-loss section later says 'Math.max not used here' and that the fix is a direct percentage cap on entry price. These statements actively contradict each other about how stop-loss management works, which can mislead users about the actual trading logic and risk controls.

Vague Triggers

Medium
Confidence
92% confidence
Finding
The invocation examples are broad enough to encourage autonomous execution such as scanning a watchlist and trading all signals without explicit confirmation boundaries. In an agent setting, vague triggers can cause overbroad actions, repeated writes, or unintended strategy execution beyond what the user meant.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill states that it maintains portfolio, journal, and observation files, but it does not clearly warn users that invoking the skill can automatically write to those files. In agent workflows, undocumented state changes reduce user awareness and can lead to silent persistence, corrupted local state, or accumulation of sensitive trading history.

Static analysis

No suspicious patterns detected.