Knowledge Graph - NL To Graph Query Translator

• Outcome: Pass
• Latest audit: Jun 8, 2026 at 9:05 AM UTC
• Version:
  1.0.0

Overview

This skill coherently helps generate graph database queries and does not include hidden execution, credential access, persistence, or network behavior.

Installers should treat generated Cypher/SPARQL as draft code: review it before execution, use read-only or least-privilege database roles when possible, parameterize user values, add LIMIT clauses for broad traversals, and avoid running CREATE/SET/DELETE examples on production data without explicit approval and backups.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Description-Behavior Mismatch

Medium

• Confidence: 90% confidence
• Finding: The reference guide includes CREATE, MERGE, SET, and DELETE examples even though the skill is described as translating natural-language questions into queries, which implies read-oriented behavior. In an NL-to-query setting, exposing write-capable patterns increases the chance that the agent generates or normalizes destructive or state-changing Cypher from ambiguous user input, enabling unintended database modification.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The skill explicitly promises to generate executable Cypher/SPARQL queries but does not warn that executing model-generated queries can retrieve broad or sensitive graph data, enable unintended graph traversal, or cause operational impact if write-capable endpoints are used. In this context, natural-language-to-query translation is inherently higher risk because vague user prompts can be turned into overbroad data access without human review.

Missing User Warnings

Low

• Confidence: 82% confidence
• Finding: Destructive operations are presented without warnings, approval requirements, or safe-execution guidance. This makes it easier for downstream users or an agent consuming the guide to treat DELETE or update statements as routine examples, increasing the likelihood of accidental destructive execution.

VirusTotal

65/65 vendors flagged this skill as clean.