ClawHub
Back to skill

Security audit

NOMOS Decision Hub

Security checks across malware telemetry and agentic risk

Overview

The skill is a decision-auditing and deployment aid whose reviewed file, environment, API, and schema behaviors are mostly disclosed and aligned with that purpose.

Before installing or using this skill in production, verify the external compliance claims, treat the included deployment commands as examples only, use a secret manager or orchestrator-managed secrets for API keys and database DSNs, and deploy with real authentication, authorization, tenant isolation, backups, and audit storage appropriate for high-stakes decisions.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
81% confidence
Finding
The skill declares no permissions, yet the described and statically detected behavior includes environment-variable access and writing files. That mismatch can cause users or orchestrators to grant execution under false assumptions, enabling secrets exposure from env vars or unintended local file modification without explicit consent.

Tp4

High
Category
MCP Tool Poisoning
Confidence
92% confidence
Finding
The skill is presented as a deterministic decision-analysis tool, but static analysis indicates additional behavior around exporting and modifying an OpenAPI schema and writing it to disk. This undisclosed functionality is dangerous because it expands the trust boundary: a user invoking a decision skill would not reasonably expect API-spec generation, security-scheme manipulation, or local artifact creation, which can be abused to alter interfaces, misrepresent auth configuration, or stage follow-on actions.

Vague Triggers

Medium
Confidence
78% confidence
Finding
The trigger scenarios are broad and ambiguous, covering generic topics like enterprise deployment, governance, stress testing, and compliance auditing. Overbroad triggers increase the chance the skill activates in contexts where it is not needed, causing unnecessary execution of its capabilities and exposing users to hidden side effects from the broader functionality identified by static analysis.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The deployment example instructs users to pass secrets such as an API key and database DSN directly via environment variables without any warning about secret management, shell history exposure, or least-privilege handling. In an enterprise-focused skill, this can normalize insecure deployment practices and lead to credential leakage or overexposed database access.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.