Security audit

libtv-cli

Security checks across malware telemetry and agentic risk

Overview

The skill appears to be a coherent LibTV CLI helper, but its installer path can download and run remote code without integrity verification.

Install only if you trust the LibTV distribution endpoints and are comfortable with a remote binary installer that does not show checksum or signature verification. Prefer running the bundled local installer over piping remote scripts, pin versions when possible, and verify the active account/project before uploads, deletes, or generation runs.

SkillSpector

By NVIDIA

Vulnerability Patterns

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
Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection

Findings (10)

Description-Behavior Mismatch

Medium

Confidence: 93% confidence
Finding: The batch installer falls back to fetching configuration from a remote API, extracting a PowerShell URL from that response, downloading the referenced script, and executing it with ExecutionPolicy Bypass. This creates a trust-on-first-use remote code execution path with no integrity pinning, signature verification, or fixed trusted installer source, so compromise of the API, returned JSON, or target hosting can lead to arbitrary code execution on the user's machine.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The document states that `account info` and `account use` recalibrate the local active-account scope and that later project commands will follow that scope, but it does not present this as a prominent warning or call out the risk of unintended operations against a different personal or team workspace. In a CLI skill that users may follow step-by-step, this can lead to accidental listing, creation, or modification of projects in the wrong account context, especially in multi-team environments.

Missing User Warnings

Medium

Confidence: 87% confidence
Finding: The documentation instructs users to upload local files into a remote project but does not explicitly warn that file contents will be transmitted to an external service and may contain sensitive data. In a CLI skill focused on project and asset management, this omission can cause unintended disclosure of private files, especially because examples encourage direct use of local paths and piping into further workflow steps.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The document exposes an internet-enabled option (`search_enabled`) and states it is enabled by default, but provides no warning that prompts or related inputs may be transmitted to external services. In a CLI skill used to operate media-generation workflows, this can lead users to unknowingly send sensitive project content, prompts, or metadata off-system, creating a real privacy and data-governance risk.

Missing User Warnings

High

Confidence: 97% confidence
Finding: The script downloads a PowerShell file to a temporary path and immediately executes it without any user-facing warning, consent, or validation. In a skill advertised as a CLI/manual for operating LibTV, silently introducing remote code execution during installation increases risk because users may not expect the installer to delegate trust to arbitrary server-provided script content.

Missing User Warnings

Medium

Confidence: 98% confidence
Finding: The document explicitly instructs users to pipe a remotely fetched PowerShell script directly into Invoke-Expression, which executes untrusted network content without prior inspection, signature verification, or checksum validation. If the hosting endpoint, CDN, API-supplied URL, or transport path is compromised, users would immediately run attacker-controlled code on their machines.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The replacement instructions tell users to recursively delete a user-supplied skill directory with rm -rf / Remove-Item -Recurse -Force and then recreate it, without strong validation, backup guidance, or safeguards against an empty or mistyped path. A user mistake, variable expansion issue, or automation misuse could cause destructive local data loss beyond the intended skill contents.

External Transmission

Medium

Category: Data Exfiltration
Content: fetch_url() { # 下载到已有路径（由调用方 mktemp）；优先 curl，其次 wget local url="$1" out="$2" if command -v curl >/dev/null 2>&1; then curl -fsSL --connect-timeout 30 --retry 2 -o "$out" "$url" elif command -v wget >/dev/null 2>&1; then wget -q -O "$out" "$url"
Confidence: 96% confidence
Finding: curl >/dev/null 2>&1; then curl -fsSL --connect-timeout 30 --retry 2 -o "$out" "$url" elif command -v wget >/dev/null 2>&1; then wget -q -O "$out" "$url" else die "need curl or wget to

External Script Fetching

High

Category: Supply Chain
Content: | 字段 | 含义 | | -------------------- | ------------------------------------------------------------ | | `version` | 最新 CLI 版本号 | | `install.shell` | macOS / Linux 一键安装脚本 URL（`curl -fsSL <url> \| bash`） | | `install.PowerShell` | Windows PowerShell 一键安装脚本 URL | | `install.CMD` | Windows cmd / 双击运行的 `.bat` URL | | `skill` | 该版本 skill 包（含本文档 + install 脚本）zip 的下载 URL |
Confidence: 97% confidence
Finding: curl -fsSL <url> \| bash

Tool Parameter Abuse

High

Category: Tool Misuse
Content: } install_from_zip_path() { # 解压到 mktemp 目录（脚本不 rm，目录会留在 $TMPDIR）；包内任一层找 libtv / libtv.exe local zip_path="$1" extract_dir binary_src [[ -f "$zip_path" ]] || die "not a zip file: $zip_path" need_cmd unzip
Confidence: 84% confidence
Finding: rm，目录会留在 $TMPDIR）；包内任一层找 libtv /

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.