ClawHub
Back to skill

Security audit

LYGO Champion: sephrael echo walker

Security checks across malware telemetry and agentic risk

Overview

The available evidence does not substantiate the advisory concern, and there is no telemetry or artifact-backed indication of malware, persistence, exfiltration, or destructive behavior.

Before installing, read the skill text for any instructions to inspect local files such as canon or verifier metadata and confirm that matches what you want. If it only provides advisor-style guidance and requires user-directed file access, the risk appears limited.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Tp4

High
Category
MCP Tool Poisoning
Confidence
73% confidence
Finding
The skill is presented as a deprecated persona helper with a 'pure advisor' stance, but the analyzed behavior suggests verification-style operations over local files and extraction/display of metadata such as champion identifiers and hash values. That mismatch can mislead users and reviewers about what the skill actually does, weakening informed consent and potentially enabling unintended local file inspection or disclosure if an implementation follows the described references.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.