Hudl AI Openclaw Model Switch

• Outcome: Review
• Latest audit: May 29, 2026 · 8:21 PM UTC
• Version: 1.0.3

Overview

This skill openly changes your OpenClaw model settings and restarts OpenClaw; the main caution is that some model aliases are broad and could pick an unintended model.

Install this only if you want chat requests to change your hudl-backed OpenClaw model and restart OpenClaw. Review the resolved model before accepting vague aliases, keep a config backup if rollback matters, and use the ClawHub install path unless you trust the GitHub fallback source.

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 (5)

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The README explicitly says the skill edits OpenClaw configuration and restarts the agent automatically, but the normal usage description does not prominently warn that model switching changes local config and causes service interruption. In an agentic environment, that can lead to unexpected persistence and downtime from a simple natural-language request, which is a real safety issue even if not overtly malicious.

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The trigger conditions are very broad and include generic phrases like 'use claude', 'smart model', 'cheap model', and 'current model', which can cause the skill to activate when a user is merely discussing models rather than authorizing a configuration change. In this skill's context, unintended invocation is more dangerous because execution can mutate local configuration and initiate a restart, creating integrity and availability risk from ambiguous natural-language input.

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The alias set for `hudl/claude-opus-4.6` includes broad descriptive phrases such as `claude`, `smart model`, `advanced`, and `big model`, which are not uniquely tied to a single model. In a model-switching skill, these vague triggers can cause unintended routing to a more expensive or different-capability model than the user intended, especially when users ask general questions about model families rather than requesting a specific switch.

Vague Triggers

Medium

• Confidence: 89% confidence
• Finding: Aliases like `flash` and `flash lite` are highly generic and overlap with common naming patterns across providers, making intent resolution ambiguous. In this skill's context, where natural-language mentions trigger model switching, these terms could select the wrong Gemini variant or collide with other 'fast' model phrasing, causing unintended configuration changes and reliability or cost issues.

Vague Triggers

Medium

• Confidence: 96% confidence
• Finding: The aliases `cheap model`, `fast model`, and `default` are context-dependent descriptors rather than unique identifiers, so they can be triggered by ordinary user language unrelated to a precise model choice. Because this skill is specifically designed to act on vague model-change phrasing, these aliases materially increase the chance of silent misselection, including downgrades in capability or unintended provider/model switching.

VirusTotal

64/64 vendors flagged this skill as clean.

Static analysis

Detected: suspicious.destructive_delete_command

Documentation contains a destructive delete command without an explicit confirmation gate.

Warn

• Code: suspicious.destructive_delete_command
• Location: README.md:90