Sales Playbook

• Outcome: Pass
• Latest audit: May 28, 2026, 12:52 PM
• Version:
  1.0.0

Overview

This is an instruction-only B2B sales playbook with no code execution, persistence, credential access, or hidden data flows, though users should handle sales and call data carefully.

Reasonable to install if you want a sales-coaching playbook. Treat customer details, pipeline data, forecasts, and call recordings as confidential: get required consent, redact unnecessary personal information, and review generated outreach for legal, privacy, and platform compliance before using it.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (7)

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: The trigger "Build my ICP" is short, natural phrasing that can plausibly appear in ordinary conversation with a general-purpose assistant. In a skill system that auto-invokes on loose phrase matching, this can cause unintended activation and route user context into the sales playbook without clear consent or scope boundaries.

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: The trigger "Write outreach for [company]" is broad and overlaps with common user requests for drafting help. This increases the chance of accidental skill invocation, which may unexpectedly apply sales-specific logic, use sensitive company context, or bypass the user's intended workflow.

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: The trigger "Prep me for discovery with [company]" is ambiguous because "discovery" is a common business term and not uniquely tied to this skill. In environments with automatic matching, the skill could activate unintentionally and consume or generate content based on sensitive prospect information.

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The trigger "Qualify this deal" is highly generic and could match many normal CRM, finance, procurement, or conversational contexts. Because the skill is designed to assess deals and influence forecasting or prioritization, accidental invocation can lead to unintended handling of potentially sensitive revenue and customer data.

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The trigger "Review my pipeline" is common phrasing in sales management and could easily be said in ordinary assistant interactions. If the platform auto-invokes the skill on this phrase, the assistant may process confidential pipeline, forecast, or account data without a sufficiently explicit tool-use boundary.

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The trigger "Forecast this quarter" is generic business language that may collide with ordinary requests for planning or projection help. In a shared or enterprise environment, this broad trigger can cause unintended activation and expose sensitive sales forecasting context to the skill path when the user did not clearly opt in.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The command "Coach me on [call recording/notes]" explicitly invites analysis of call recordings and notes, which commonly contain personal data, confidential customer information, or regulated content. Without a privacy warning, consent guidance, or handling restrictions, users may upload sensitive material without understanding retention, sharing, or compliance implications.

VirusTotal

64/64 vendors flagged this skill as clean.