Security audit

LocalMiniDrama短剧助手

Security checks across malware telemetry and agentic risk

Overview

This is a coherent LocalMiniDrama connector, but users should only point it at a backend they trust because it can change projects and send scripts, media, and API keys there.

Install this only if you operate or trust the configured LocalMiniDrama backend. Prefer localhost or controlled infrastructure, avoid public tunnels for private content or keys, and manually confirm delete, import/export, bulk generation, and API-key update actions before allowing the agent to run them.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (6)

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The README advertises very broad natural-language triggers such as asking to create, generate, export, import, configure, or check status, and these overlap with ordinary conversation. In an agent environment, this increases the chance of unintended invocation of state-changing actions like project creation, bulk generation, export, or configuration changes without sufficiently explicit user intent.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The documentation describes capabilities that can alter system state or handle data—such as export/import, AI configuration, and batch media generation—but does not warn users about cost, privacy, overwrite, or external-network implications. In this skill context, those actions may expose project data, consume compute resources, or modify backend configuration if triggered too easily.

Vague Triggers

Medium

Confidence: 81% confidence
Finding: The trigger condition is described broadly as any request related to creating, managing, or generating short-drama content, without tight scoping or exclusion rules. In an agent environment, this can cause the skill to activate on loosely related user prompts and perform networked or state-changing actions unexpectedly.

Vague Triggers

High

Confidence: 92% confidence
Finding: The trigger list contains many broad natural-language phrases such as '生成视频', '生成图片', '配置API', and '查看项目' without disambiguation. This creates high semantic-collision risk where unrelated conversations may invoke a skill capable of destructive, expensive, or sensitive backend operations.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The skill exposes high-impact operations including deletion, import/export, batch media generation, and AI key configuration, but the instructions do not require explicit user confirmation or safety interlocks before executing them. In an agent setting, that increases the risk of accidental destructive changes, sensitive data handling, and unwanted spend through one-shot natural-language activation.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The trigger list contains many generic phrases such as '生成视频', '生成图片', '查看项目', and '配置API' that are likely to appear in normal user conversations unrelated to this specific skill. This can cause unintended invocation of a skill that has HTTP and memory access, increasing the chance of accidental backend requests, state changes, or project/configuration operations without clear user intent.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal