ClawHub
Back to skill

Security audit

LYGO Public Witness

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed public-feed witness tool with narrow network use and no evidence of hidden publishing, credential access, or destructive behavior.

Installers should treat this as a public-data CLI: expect outbound HTTPS GETs to the listed public sources, use report-writing flags only with paths you intend to create, and run the Ollama command only if you want a local Ollama service to receive and summarize the fixed witness prompt.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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 (1)

Missing User Warnings

Low
Confidence
88% confidence
Finding
The ollama subcommand sends a prompt to a local HTTP service on 127.0.0.1:11434 without any runtime disclosure or confirmation, which can surprise users and forward locally available context to another process. Although the destination is localhost and the feature is described as optional in comments/docstrings, the execution path itself does not present an explicit user-facing warning at the point of use.

Static analysis

Detected: suspicious.install_untrusted_source

Install source points to URL shortener or raw IP.

Warn
Code
suspicious.install_untrusted_source
Location
claw.json:44