Code Graph(代码图谱)

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed GitNexus setup skill, but users should review the environment-changing setup and git troubleshooting commands before running them.

Install this only if you want GitNexus added to your environment and your editor/MCP configuration updated. Review npm/npx commands before execution, and do not run the git add/commit troubleshooting example unless you have reviewed .gitignore and the files that would be committed.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to run commands that can globally install software and modify editor/MCP configuration without requiring an explicit warning or confirmation from the user at the point of action. This is risky because these changes affect the user's broader environment beyond the current project, may alter trusted tooling behavior, and are not easily scoped or undone if triggered unintentionally.

Missing User Warnings

High
Confidence
98% confidence
Finding
The troubleshooting section suggests `git init && git add . && git commit -m "initial"` without a strong warning that this creates a repository and commits the entire current directory contents. That can accidentally capture secrets, large/generated files, or unrelated data and permanently change repository state, which is especially dangerous in an automated skill context.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal