ClawHub
Back to skill

Security audit

Longbridge

Security checks for vulnerabilities and agentic risk

Overview

This skill is mostly a read-only Longbridge connector, but its setup path tells an agent to run remote installer scripts directly in the shell, which needs careful review.

Install only if you trust OOMOL's Longbridge connector and understand it can read private brokerage/account information. Do not let an agent run the pipe-to-shell installer automatically; install the `oo` CLI through a reviewed official method, and separately confirm any action that could ever write, trade, or change account state.

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
  • 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)

External Script Fetching

High
Category
Supply Chain
Confidence
93% confidence
Finding

The skill instructs the agent to install software via a remote script piped directly into the shell (curl ... | bash), which is a classic supply-chain and arbitrary code execution risk. In this context, the command would run with the user's local privileges and could be triggered from a troubleshooting path inside a finance-related skill, increasing the chance of unsafe execution on a user machine.

Content

Scanner excerpt · SKILL.md (reported line 134)May include surrounding context.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest description frames the skill as only for "searching and reading data," but the documented actions expose sensitive brokerage/account information such as cash balances, positions, orders, executions, watchlists, and profit analysis. This mismatch can mislead users or higher-level agents into treating the skill as low-risk market-data access when it can retrieve private financial data tied to the connected account.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The safety model says state-changing actions will be tagged as [write] or [destructive], but the action list contains no such tags despite exposing trading- and account-adjacent capabilities and instructing the agent to trust the live schema. If any write-capable action exists now or is added later without documentation parity, an agent could execute sensitive operations without the promised confirmation guardrail.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.