Security audit

Research Assistant

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed research workflow that saves notes in a local research folder, with no executable code or hidden data access found.

Install this only if you want your agent to maintain a persistent research/ folder. Keep monitoring disabled unless you intentionally configure it, review generated briefs before relying on them, and avoid storing confidential research topics in shared or indexed workspaces.

SkillSpector

By NVIDIA

Vulnerability Patterns

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
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The skill directs the agent to create a research directory, save briefs, and automatically update `research/README.md` without a clear upfront warning that local files will be created and modified. In environments where agents have filesystem access, this can lead to unexpected persistent writes, workspace clutter, disclosure of sensitive research topics in indexed files, or unintended modification of repository contents.

Missing User Warnings

Low

Confidence: 87% confidence
Finding: The file explicitly instructs the agent to 'automatically create a new research brief' when monitoring detects a significant event, but it does not require user confirmation, write-scope limits, or a visible notice that files will be created. In an agent skill, implicit file writes can surprise users, create clutter, and be abused by noisy or manipulated triggers to generate unwanted artifacts over time.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal