Security audit

Db Schema Designer Free

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed SQLite schema-design helper with local write and command examples that fit its purpose, but users should confirm paths before allowing any database changes.

Install only if you want an agent to help design and optionally create local SQLite schemas. Before allowing write or exec actions, confirm the target project path, database file, and SQL statements, and avoid using callback_url unless you explicitly trust the destination and know what will be sent.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Behavioral ASTexec() Call, eval() Call, Dynamic Import
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (5)

Description-Behavior Mismatch

Medium

Confidence: 92% confidence
Finding: The skill is presented as a schema design assistant, but its instructions extend into operational database actions such as create/query/export. That scope expansion can cause an agent to perform real modifications instead of only generating design guidance, increasing the chance of unintended state changes in local data stores.

Description-Behavior Mismatch

Medium

Confidence: 95% confidence
Finding: These instructions tell the agent to create directories and run sqlite3 commands in a user project, which goes beyond passive schema advice into filesystem and database modification. In an agent environment with write/exec tools, that creates a real risk of unintended local changes or execution in the wrong project context.

Context-Inappropriate Capability

Medium

Confidence: 87% confidence
Finding: A callback_url parameter introduces an outbound communication path that is not necessary for a local SQLite schema design workflow. In practice, this could be abused to exfiltrate generated data, metadata, or execution status to an attacker-controlled endpoint if an agent blindly honors the parameter.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The trigger condition is broad enough to activate on many generic database requests, which increases the chance that the skill will be invoked outside its intended context. In an environment with write and exec permissions, overbroad routing can lead to inappropriate command suggestions or unintended database/file modifications.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The skill includes guidance that can write files and execute commands without clearly warning that local project directories and database contents may be modified. Lack of explicit warning and consent is dangerous in agentic environments because users may expect design advice but receive side effects instead.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.