Jarvis Db Query Doctor 01

• Outcome: Pass
• Latest audit: May 28, 2026, 12:54 PM
• Version:
  1.0.0

Overview

This is an instruction-only planning checklist that is mislabeled as a SQL query doctor, but it does not request credentials, code execution, persistence, or data access.

Install only if you want a generic planning and quality-gate checklist. If you specifically need SQL performance help, expect to provide the query, schema, indexes, and execution plan yourself, and consider asking the publisher to narrow the skill text and examples to database work.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Description-Behavior Mismatch

High

• Confidence: 97% confidence
• Finding: The skill metadata claims a narrow SQL diagnosis/query-rewrite purpose, but the actual body implements a generic planning workflow. This mismatch can cause the agent to invoke the skill for broad tasks outside its declared safety and capability boundaries, leading to unintended behavior, overbroad access to user tasks, and bypass of domain-specific review expectations.

Intent-Code Divergence

Medium

• Confidence: 94% confidence
• Finding: The example prompts present the skill as a general planning assistant rather than a SQL-performance specialist, directly contradicting the stated purpose. In practice, examples strongly influence routing and use, so these broad prompts increase the chance of accidental or inappropriate invocation on unrelated work.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The description uses broad language about development work, repeatable steps, and practical outputs, which can match many ordinary requests unrelated to SQL. Overly broad invocation criteria make the skill easier to trigger in contexts where its instructions are not appropriate, increasing misrouting risk and weakening least-privilege skill selection.

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The body defines a generic execution workflow with no database-specific constraints, triggers, or prohibitions, so it functions as a catch-all planning skill. In an agent environment, this broadness can let the skill capture unrelated tasks and produce outputs beyond its declared domain, undermining predictability and safety controls.

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: The example prompts are broad enough to match common workplace requests like turning notes into a plan or producing an action sequence, which are unrelated to SQL diagnosis. Because examples heavily shape model behavior, these prompts materially increase the chance that the skill will be selected as a general-purpose assistant rather than a scoped database tool.

VirusTotal

66/66 vendors flagged this skill as clean.