sales-powermap

• Outcome: Review
• Latest audit: May 28, 2026, 1:51 AM
• Version:
  1.0.0

Overview

This sales lead-generation skill is transparent about its purpose, but it directs agents to collect and profile named people’s business contact and social information with weak privacy boundaries.

Install only if you are comfortable with a skill that builds person-level sales prospect lists. Use it only for compliant B2B outreach, restrict it to publicly available business contact data, avoid personal emails/phones and speculative email inference, and review all generated leads and messages before use.

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 (4)

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The trigger conditions are broad enough to activate on common sales-related requests, which increases the chance the skill runs without clear user intent for deep prospecting or contact discovery. In this skill, accidental activation is more dangerous because the workflow proceeds quickly into company targeting, org-chart mining, and contact collection with minimal confirmation.

Missing User Warnings

High

• Confidence: 98% confidence
• Finding: The skill explicitly includes '联系方式获取' and contact-detail collection as a core function, but provides no privacy notice, lawful-use boundary, or data-handling constraints. In context, this materially increases risk because the workflow seeks personal profiles, social accounts, and inferred emails for named individuals and packages them into outreach-ready reports.

Ssd 3

High

• Confidence: 99% confidence
• Finding: These instructions direct the agent to gather and output individuals' contact details, including inferred corporate email patterns and social profiles, which enables targeted unsolicited outreach and identity-based profiling at scale. The risk is heightened by the skill's explicit recommendation logic for the 'best channel' per person, making the output immediately operational for intrusive or abusive campaigns.

Ssd 4

Medium

• Confidence: 95% confidence
• Finding: The workflow systematically moves from identifying target companies to mapping internal hierarchies, selecting vulnerable entry points, and designing multi-step '包抄' outreach paths against named individuals. Even if framed as sales enablement, this structure can be repurposed for social engineering, harassment, or influence operations because it operationalizes relationship intelligence and approach sequencing.

VirusTotal

VirusTotal findings are pending for this skill version.