Detectify

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Detectify integration that uses Membrane to connect to a Detectify account and can perform both read-only and account-changing actions.

Install this only if you trust Membrane to broker access to your Detectify account. Use the least-privileged Detectify account practical, review the browser authorization flow, and require clear confirmation before deleting assets or scan profiles, stopping scans, or changing schedules/settings.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
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 (2)

Context-Inappropriate Capability

Medium

Confidence: 92% confidence
Finding: The documented generic proxy allows arbitrary requests to the Detectify API, which materially expands the skill's effective permissions beyond the narrowly described purpose. That broad surface can enable unintended read/write/destructive operations if an agent follows user prompts too loosely or if the skill is invoked under an over-broad trust assumption.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill advertises destructive operations such as deleting assets/scan profiles and stopping scans without any guidance to confirm intent, display consequences, or require a second step. In an agent setting, this increases the chance of accidental or prompt-induced destructive actions against production security-scanning configuration.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal