Security audit

HR简历分析评分报告

Security checks across malware telemetry and agentic risk

Overview

This resume-screening skill appears useful, but it goes beyond basic resume matching into sensitive candidate judgments, verification-record collection, and local storage of candidate data without enough safeguards.

Install only if you are prepared to handle candidate data under your recruiting, privacy, and employment-law obligations. Use it for objective extraction and JD matching, avoid relying on personality/authenticity/risk labels for decisions, get explicit consent before collecting verification records, and set clear retention/deletion rules for any local reports it writes.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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

Findings (5)

Context-Inappropriate Capability

Medium

Confidence: 93% confidence
Finding: The skill expands from resume parsing and JD matching into personality profiling, leadership-potential inference, authenticity scoring, and risk classification. These are high-stakes employment inferences about candidates and can enable unfair or non-transparent decision-making, especially when presented as automated assessments without clear validation, consent, or human-review safeguards.

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The skill recommends collecting third-party verification materials such as academic certificates, separation proofs, and social security records, which materially increases the sensitivity of data being handled. This broadens data collection beyond the stated scope and creates privacy, retention, and misuse risks if users are encouraged to gather and store highly sensitive employment and identity records unnecessarily.

Vague Triggers

Medium

Confidence: 86% confidence
Finding: The manifest description uses broad trigger language around resume screening, scoring, and report generation, which can overlap with ordinary recruiting conversations. Over-broad activation can cause the skill to engage in contexts where users did not intend file parsing, scoring, or report generation, increasing the chance of processing sensitive candidate data unexpectedly.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: Several example trigger phrases are short and generic, such as requests to score resumes or filter candidates, without strong scope constraints. In a recruiting assistant context, this makes accidental invocation more likely and may lead to unintended handling of personal data or premature candidate evaluation.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The skill instructs saving pasted resume text to local files and generating Excel/JSON/Markdown reports containing candidate details, but does not clearly warn users that sensitive personal information will be written to disk. Because resumes commonly contain phone numbers, email addresses, employment history, and education data, silent local persistence materially raises privacy and data-exposure risk.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.