ClawHub
Back to skill

Security audit

Local LLM Setup Advisor

Security checks for vulnerabilities and agentic risk

Overview

The skill provides local AI setup guidance, with overly broad activation keywords but no material security concern in the inspected files.

Use this skill for local LLM setup and planning. Its broad keywords may activate it for unrelated questions; review generated commands for suitability before running them.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Broad trigger keywords can activate the skill for requests unrelated to local LLM setup.

Low
Category
Trigger Abuse
Confidence
97% confidence
Finding

The trigger list includes privacy and software-and-data without requiring a local LLM context. These terms cover many unrelated requests, and the skill provides no exclusions or explicit contextual gate to prevent unintended activation.

Content

No source excerpt is available for this finding.

Broad activation keywords can invoke the local-LLM skill for unrelated requests.

Low
Category
Trigger Abuse
Confidence
97% confidence
Finding

The trigger list includes privacy and software-and-data, which cover many requests unrelated to setting up local LLMs. It does not require these keywords to appear in a local-inference context or provide exclusions, so ordinary privacy or software questions could unintentionally activate this skill.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.