ClawHub
Back to skill

Security audit

Feature Retcon / 需求后悔药

Security checks across malware telemetry and agentic risk

Overview

This skill is a local, user-confirmed workflow for revising project artifacts with recovery logging, and I found no hidden exfiltration or destructive behavior.

Before installing, understand that this skill can modify and delete files in roots you explicitly authorize after its assessment phase. Do not authorize sensitive files unless you accept that local recovery payloads may contain their original contents, even though the contract is kept at 0600 and should not be committed.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (1)

Lp3

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding
The skill instructs the agent to use shell commands, read files, and inspect environment/project state, but it does not declare corresponding permissions or capability expectations. This creates a transparency and policy-enforcement gap: operators may approve or invoke the skill without realizing it can access local files and execute commands, increasing the chance of unintended data exposure or unsafe command execution in a sensitive repository.

VirusTotal

58/58 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.