Security audit

Portable Tools

Security checks across malware telemetry and agentic risk

Overview

This is a legitimate portability guide, but it repeatedly encourages users to expose real OAuth/keychain token values during debugging.

Review before installing or following the examples. The checklist script itself looks limited, but users should not paste, print, screenshot, or log real access tokens, refresh tokens, keychain password output, or raw credential JSON. Use redacted prefixes/suffixes, lengths, hashes, field-presence checks, and expiry timestamps instead.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 88% confidence
Finding: The skill contains shell command examples that access sensitive system stores and files, but it declares no permissions or equivalent warning/metadata. That mismatch can cause downstream systems or users to underestimate what the skill may prompt them to run, increasing the chance of unsafe execution or review bypass.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The documentation shows token-like values in BEFORE/AFTER examples and frames success as proving token changes with exact values. Even if illustrative, this normalizes copying, exposing, or logging secrets into chat, terminals, screenshots, or docs where they may be retained or leaked.

Missing User Warnings

High

Confidence: 98% confidence
Finding: These debugging instructions explicitly ask the user to reveal exact values and identify which keychain entries contain tokens, without any safeguard about redaction. This creates a direct credential disclosure risk because users may paste live secrets from password stores into logs, chats, or issue trackers.

Missing User Warnings

High

Confidence: 97% confidence
Finding: The real-world example retrieves keychain data, validates it, and logs account selection/errors around credential-bearing content, but the surrounding documentation provides no warning against exposing secrets during troubleshooting. In this context, readers are encouraged to operate directly on live credential stores and may inadvertently print or share sensitive data.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.exposed_secret_literal, suspicious.generated_source_template_injection

File appears to expose a hardcoded API secret or token.

Critical

Code: suspicious.exposed_secret_literal
Location: SKILL.md:51

User-controlled placeholder is embedded directly into generated source code.

Critical

Code: suspicious.generated_source_template_injection
Location: SKILL.md:132