ClawHub
Back to skill

Security audit

Web3 AI Agent项目专业筛选器

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only Chinese Web3 AI project screening guide that uses web search for investment research and does not install code, persist itself, or access private data.

Install only if you want a Chinese-language framework for researching Web3 AI Agent investments. Treat its ratings and buy-window suggestions as research prompts, not financial advice, and verify market data from current authoritative sources before making decisions.

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
  • 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 (2)

Natural-Language Policy Violations

Medium
Confidence
88% confidence
Finding
The skill is written to operate in Chinese and defines output/workflow entirely in Chinese without offering a user-language choice. This can cause misunderstanding of financial risk disclosures, evaluation criteria, or recommendations when invoked for users expecting another language, which is especially problematic in an investment-screening context.

Vague Triggers

Medium
Confidence
94% confidence
Finding
The skill’s activation guidance is broad enough that it could trigger on ordinary project-related prompts, causing the agent to invoke this investment-oriented workflow when the user did not explicitly ask for it. Over-broad activation increases the chance of unintended tool use, irrelevant financial analysis, and policy-bypassing context capture from benign requests.

Static analysis

No suspicious patterns detected.