ClawHub

Supernal Interface CLI

Security checks across malware telemetry and agentic risk

Overview

This appears to be a legitimate developer CLI skill, but it gives an agent broad authority to install software, rewrite project files, change IDE/MCP configuration, and install other agent components without strong safety guidance.

Review the @supernal/interface package and publisher before installing. Run the CLI first in a disposable project or clean git branch, prefer --dry-run and --manual modes, avoid --force and --git-commit until you have reviewed diffs, and do not allow an autonomous agent to run setup-mcp or setup-claude without explicit approval.

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
86% confidence
Finding
The documentation advertises `setup-mcp --force` as an overwrite operation and describes the setup as 'fully automated' with 'zero manual steps', but it does not warn that existing IDE/MCP configuration or related files may be replaced. In an agent skill context, terse command examples can be copied or invoked automatically, so omission of a clear warning increases the chance of accidental destructive changes.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
This section documents commands such as `si init --inject`, `si init --migrate`, `si init --revert`, and `si migrate-routes` that modify source files, but it does not clearly warn users that these commands rewrite project code and imports. In a skill consumed by agents, file-modifying commands without prominent change-control guidance can lead to unintended source edits or broad refactors being run without adequate review.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal