ClawHub
Back to skill

Security audit

HyperGrok Desk Trade Lifecycle

Security checks for vulnerabilities and agentic risk

Overview

This is a clearly scoped trading workflow skill with strong approval and journaling controls, though it applies to high-impact financial actions.

Install only if you intend agents to help manage a trading workflow. Confirm your broader desk configuration keeps exchange keys, approval gates, and mainnet execution under controls you trust, and use testnet first when possible.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (1)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 98)May include surrounding context.

md
Each adjustment receives fresh evidence and a new Risk Manager PASS, the exact suffixed ticket id for user approval, and its own expiry. Earlier approval or Risk PASS does not carry forward. Closing a position at market needs a ticket stating the reduce-only size read live from the account and the slippage bound. Execute once within expiry, then reconcile the adjustment against the exchange record under the parent proposal.

The only exception is a pre-authorised protective action the user has written into `desk.md` (for example "the Execution Trader may cancel orphaned stops after a position closes without asking"). Even then, the action is journaled.

## Paper trading and testnet