Security audit

Threads User Posts

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed Threads profile-post scraper, but users should treat its login, network-response reading, and batch-collection guidance carefully.

Install only if you intend to use browser automation to collect Threads profile posts. Use it on public or otherwise authorized profiles, do not let it capture credentials, review any batch script before running it, and avoid large-scale or parallel scraping that could violate platform rules or user privacy expectations.

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 (4)

Context-Inappropriate Capability

Medium

Confidence: 94% confidence
Finding: The skill goes beyond retrieving public profile content by explicitly assisting with Threads login flow. That expands scope from passive scraping of public data into handling authenticated sessions, which can expose credentials, session state, or private/account-scoped content if misused. In this context, the broader capability makes the skill materially more sensitive than its stated purpose.

Context-Inappropriate Capability

Low

Confidence: 89% confidence
Finding: The skill instructs inspecting network requests and GraphQL responses, which broadens access from visible page content to backend-delivered data. Even if used for pagination, network inspection can reveal metadata or response fields not intended for direct user-facing extraction, increasing the chance of overcollection or access to account-scoped information in an authenticated session.

Context-Inappropriate Capability

Low

Confidence: 91% confidence
Finding: The batch orchestration and multi-session guidance turns a single-profile extraction skill into a scalable scraping workflow. That increases abuse potential, operational footprint, and the likelihood of violating platform controls or causing rate-limit evasion behavior, especially when paired with login use and pagination automation.

Missing User Warnings

Low

Confidence: 82% confidence
Finding: The skill provides automation guidance for batch collection without clearly warning the user about rate limiting, platform impact, or abuse risks. While not a direct exploit primitive, the omission encourages unsafe operational behavior and makes misuse easier in practice.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.