ClawHub
Back to skill

Security audit

图谱

Security checks for vulnerabilities and agentic risk

Overview

The skill appears to be a loosely scoped Graph Protocol helper, but it asks for broad read/write/command authority without clear boundaries and contains mismatched instructions.

Review this skill carefully before installing. It does not show malicious code, but its declared tool access is broader than its Graph Protocol purpose justifies. Install only if you are comfortable with a skill that may guide an agent toward file and shell actions, and restrict command execution or filesystem access where your agent platform allows it.

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
93% confidence
Finding
The declared JSON return format describes a generic grading/audit report with fields like overall_grade, code style, and security compliance, which does not match the stated purpose of routing blockchain data questions to Graph Protocol services. This mismatch can cause an agent to invoke the skill under false assumptions, mishandle outputs, or chain the result into unrelated workflows, increasing the risk of unsafe automation and incorrect downstream decisions.

Intent-Code Divergence

Medium
Confidence
91% confidence
Finding
The workflow section mixes Graph routing claims with unrelated steps for file preparation, command execution, and runtime command invocation. In a skill that declares tools including exec and write, ambiguous documentation can mislead an agent into performing filesystem or shell actions that are outside the advertised scope, expanding the attack surface and enabling unintended execution paths.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The activation wording says to use the skill for data analysis, report generation, statistical insight, and visualization, which is far broader than routing blockchain data questions to Graph Protocol services. Overbroad triggers can cause an agent to select this skill for many general analytical requests, exposing read/exec/write capabilities in contexts where they are unnecessary and potentially unsafe.

Static analysis

No suspicious patterns detected.