ClawHub
Back to skill

Security audit

JSON修复工具

Security checks for vulnerabilities and agentic risk

Overview

The skill is meant to repair JSON files, but it asks for broad command and file authority while giving under-scoped and partly unsafe instructions for executing JavaScript-based parsing.

Review this skill carefully before installing. Use it only on explicitly chosen JSON files or directories, avoid untrusted inputs, and do not rely on JavaScript VM evaluation as a safe parser. Prefer non-evaluating JSON repair tools and require confirmation before overwriting or recursively modifying files.

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 (2)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill claims comment stripping and safety properties that do not match actual JavaScript parsing behavior, and it promotes `vm.runInNewContext` as a safe way to process malformed JSON-like input. In context, this can mislead an agent or operator into feeding untrusted data into a JavaScript evaluator, increasing the risk of code execution or sandbox escape assumptions if implementation follows the documentation.

Vague Triggers

Medium
Confidence
86% confidence
Finding
The trigger language is overly broad and framed around generic productivity and automation scenarios, so an agent may invoke this skill for unrelated requests. In a skill that declares `exec`, `read`, and `write` capabilities and suggests file/directory processing, over-selection increases the chance of unnecessary command execution or modification of user files in the wrong context.

Static analysis

No suspicious patterns detected.