ClawHub
Back to skill

Security audit

Interview Analysis

Security checks across malware telemetry and agentic risk

Overview

This is a text-only interview analysis skill that creates local Markdown hiring notes; it is coherent with its stated purpose but should be used carefully because the outputs may contain sensitive candidate information.

Install only in workspaces appropriate for confidential hiring material. Review generated recommendations for bias and accuracy, restrict access to saved candidate notes, consider using pseudonymous candidate IDs in paths, and keep humans responsible for employment decisions.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (4)

Description-Behavior Mismatch

Medium
Confidence
93% confidence
Finding
The template's stated purpose and sections direct the agent to analyze interviewer methodologies and derive organizational hiring recommendations, which materially differs from the skill's declared purpose of evaluating candidate capability. This creates a spec-to-template misalignment that can cause the system to collect, infer, and output sensitive organizational assessments outside the advertised scope, increasing the risk of unintended data handling and misleading users about what the skill actually does.

Missing User Warnings

Low
Confidence
81% confidence
Finding
The README explicitly says the skill writes files into candidate-specific directories but does not warn users that local files will be created or modified. In agent environments, undisclosed write behavior can surprise users, cause unintended workspace changes, and increase risk if candidate names or paths are not carefully controlled elsewhere.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill directs the agent to save candidate analysis cards under candidate-specific paths, which can create persistent files containing interview content, evaluations, and inferred sensitive personal or employment data without any privacy warning, minimization rule, or consent check. In a hiring context, this increases the risk of unnecessary retention, accidental disclosure, and noncompliant handling of confidential candidate information.

Hidden Instructions

High
Category
Prompt Injection
Content
> [!NOTE]
> **Analysis Perspective**: Based on {Expert_Persona_1} and {Expert_Persona_2}'s hiring philosophy.
> **Core Conclusion**: A **{core determination: e.g., strong executor / disguised strategist}**, but **{qualifying assessment: e.g., lacks foundational innovation / possesses high potential}**.
<!-- 
Guidance for AI:
Dynamically select {Expert_Persona_X} based on candidate's role:
Confidence
94% confidence
Finding
<!-- Guidance for AI: Dynamically select {Expert_Persona_X} based on candidate's role: **Expert Selection Strategy**: - Persona 1 (Domain Expert): Top methodology master in the role's field - Exam

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.