HireMate — AI招聘助手

• Outcome: Review
• Latest audit: May 29, 2026 at 9:36 PM UTC
• Version:
  1.0.0

Overview

This recruiting skill is a Review item because it can score and recommend candidates using sensitive hiring criteria without enough privacy or fairness guardrails, though no malware or hidden persistence is evident.

Review before using with real applicants. Use only authorized candidate data, minimize or redact personal details, keep outputs private, and treat scores as draft decision support requiring human HR/legal review. Consider removing or constraining cultural-fit, employment-gap, default multilingual, auto-reject, and no-hire recommendations unless they are job-related, documented, and compliant for the hiring context.

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 (3)

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The trigger list includes very broad terms such as "hire," "recruit," "JD," and compensation-related phrases that can appear in many benign conversations. Overly broad activation can cause unintended invocation of a skill that processes resumes, candidate data, salary information, and file inputs, exposing sensitive HR workflows or causing the wrong tool to act on user data.

Natural-Language Policy Violations

Medium

• Confidence: 94% confidence
• Finding: The template includes 'Multilingual capabilities' as a positive screening criterion without any requirement that language ability be job-related or explicitly requested by the user. In a recruiting assistant, this can encourage discriminatory hiring practices or proxy selection based on national origin, especially when the generated criteria are used to automate or standardize resume screening.

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The script accepts a user-supplied --output path and writes to it directly without validation, restriction, or overwrite confirmation. In an agent or automation context, this can modify arbitrary local files accessible to the process, which is unsafe even though the content written is only a generated report.

VirusTotal

66/66 vendors flagged this skill as clean.