ClawHub

GTM Skill Packs

Security checks across malware telemetry and agentic risk

Overview

This GTM skill is mostly coherent, but it can direct agents to sync prospect data into external ad and outreach tools without clear approval or rollback controls.

Install only if you want a GTM assistant that can research accounts and work with connected audience or outreach tools. Before any push or sync to LinkedIn Ads, Smartlead, or similar systems, require the agent to show the exact records, fields, destination account, expected side effects, and rollback plan, then approve the action explicitly.

SkillSpector

By NVIDIA
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 (3)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The instruction to use the skill for 'any GTM task requested' creates an overly broad activation boundary, increasing the chance the agent will invoke this skill for loosely related or ambiguous requests. That can cause inappropriate tool use, unintended data handling, or the application of sales workflows where the user's intent was not specifically to run this skill pack.

Vague Triggers

Medium
Confidence
81% confidence
Finding
The trigger phrase 'who should I target' is generic and can appear in ordinary conversation outside a deliberate GTM workflow. Because it lacks domain qualifiers, the agent may activate this skill on ambiguous prompts and steer the interaction into prospecting or targeting behavior the user did not intend.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The description uses broad, natural-language trigger phrases like "qualify a prospect," "score a lead," and "decide who to call first," which are common business requests and can cause the skill to activate in ordinary conversation without clear user intent. In an agent environment, this can lead to overbroad routing, unintended use of the scoring framework, and downstream business decisions being made from an automatically-invoked skill the user did not explicitly request.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal