EngineMind

Security checks across malware telemetry and agentic risk

Overview

EngineMind is mostly a coherent local research tool, but its dashboard exposes runtime and locally derived data on all network interfaces without authentication.

Install only if you are comfortable reviewing and manually running the code. Before running the dashboard script, bind it to 127.0.0.1 or firewall port 8888, and avoid feeding private documents unless you accept that derived state and reports will be written to local logs.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (3)

Missing User Warnings

Low

Confidence: 78% confidence
Finding: The script automatically reads local progress and memory JSONL files from disk without explicit user consent, disclosure, or path restrictions beyond fixed filenames. In an agent skill context, silent ingestion of local state can expose sensitive historical data, model traces, or user-derived memory into outputs or downstream processing, especially if operators do not realize the skill accesses persistent local files.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: This code reads arbitrary file contents and feeds them into processing that immediately persists derived snapshots and source identifiers to disk, while the module is explicitly designed for continuous append-only logging. In a skill context, that creates a real privacy and data-governance risk because sensitive local file content may be ingested and recorded without user consent, clear disclosure, minimization, or safeguards around what gets stored and for how long.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The script starts a local HTTP server on all interfaces using ('', 8888) and several endpoints explicitly set Access-Control-Allow-Origin: * with no authentication or consent gate. This exposes detailed runtime state and local files/endpoints to any host that can reach the service, and the permissive CORS policy makes browser-based access from arbitrary origins easier if the service is reachable from a user's network.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal