Local Llm Setup Advisor

• Outcome: Pass
• Latest audit: Jun 14, 2026 at 7:22 AM UTC
• Version:
  0.1.1

Overview

This is a guidance-only skill for remote llama.cpp setup, with the main caveat that its auto-invocation wording is broader than its narrow purpose.

Install only if you want a lightweight planning aid for remote llama.cpp server URL endpoint support. Be aware it may be selected for broader feature-request or support wording unless the publisher narrows the triggers or disables implicit invocation.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• 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
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Vague Triggers

Medium

• Confidence: 96% confidence
• Finding: The skill description includes very broad routing terms like "work-productivity," "feature," "request," "support," and "remote," which are common across many unrelated user requests. This can cause the skill to be invoked outside its narrow purpose, increasing the chance that users receive irrelevant or misleading guidance and that this skill overrides better-matched skills in multi-skill environments.

Vague Triggers

Medium

• Confidence: 98% confidence
• Finding: The Triggers section lists generic keywords such as "feature," "request," "support," and "remote" without contextual constraints or negative examples. In a skill-routing system, this broad match surface can lead to frequent accidental invocation, creating prompt-routing confusion and potentially exposing users to incorrect workflows for unrelated tasks.

Vague Triggers

Medium

• Confidence: 96% confidence
• Finding: The default prompt and description use very broad trigger terms such as 'work-productivity, feature, request, support, remote' plus generic deliverables like 'checklist, analysis, or implementation support.' With implicit invocation enabled, this can cause the skill to be auto-selected for many unrelated requests, expanding its influence beyond the narrow llama.cpp remote endpoint use case and increasing the chance of prompt hijacking, irrelevant activation, or unintended handling of sensitive workflows.

VirusTotal

64/64 vendors flagged this skill as clean.