ClawHub
Back to skill

Security audit

投资研究

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent Cue investment-research wrapper, but it asks the agent to fetch mutable remote code and run it locally with access to a Cue API key and paid credits.

Review the Cue runner source and publisher before installing, and prefer a pinned, audited checkout if you use it. Expect the skill to use network access, read your Cue API key from ~/.cue/config.json, consume credits after confirmation, and store/update code under ~/.cue/cue-skills.

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 (3)

Natural-Language Policy Violations

Medium
Confidence
95% confidence
Finding
The manifest description and the entire user-facing instructions are written in Chinese, and the skill name/usage text does not indicate any option to choose another language. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is explicitly documented and justified.

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
The skill directs the agent to clone or update a remote Git repository and then rely on code from that repository as part of normal operation. That expands the skill from documentable research guidance into arbitrary supply-chain code retrieval and execution, creating a path for malicious repository changes, mirror compromise, or unexpected local side effects on the host.

Context-Inappropriate Capability

Medium
Confidence
96% confidence
Finding
The skill requires executing a local Python script to fulfill an investment-research request, which grants code-execution capability beyond the stated business purpose. Even if the script is intended as a legitimate runner, it can access local files, network resources, and user environment secrets, so the skill effectively becomes an execution vector rather than a bounded research workflow.

Static analysis

No suspicious patterns detected.