Generator Bangla

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed cloud video-generation skill, but uploaded media and prompts are processed by NemoVideo and may use credits.

Install only if you are comfortable sending videos, audio, images, text prompts, and related metadata to NemoVideo's cloud service. Use a limited-purpose NEMO_TOKEN where possible, monitor credit usage, and avoid confidential or regulated media unless you trust the provider's data handling.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (6)

Description-Behavior Mismatch

Medium

Confidence: 91% confidence
Finding: The manifest markets a narrowly scoped Bangla caption/voiceover generator, but the instructions expose a substantially broader remote video editing and rendering surface, including uploads, state inspection, SSE-driven edits, and export workflows. This scope mismatch can mislead users and host systems about what the skill is permitted to do, increasing the chance of unintended file handling, remote processing, or policy bypass through a seemingly specialized skill.

Context-Inappropriate Capability

Medium

Confidence: 93% confidence
Finding: The skill is instructed to obtain anonymous tokens, create backend sessions, and manage credits, which goes beyond a simple media-processing interface and introduces account-like actions without clear user consent. That expands the trust boundary to authentication and billing-related flows, creating risk of unauthorized resource consumption, opaque third-party account linkage, or abuse of free-credit mechanisms.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The starter phrase "generate my video clips or text" is so broad that ordinary conversation could accidentally trigger the skill. In a skill that uploads media and contacts a remote backend, overbroad activation raises the risk of unintended invocation, privacy exposure, and confusing handoff into external processing.

Vague Triggers

Medium

Confidence: 87% confidence
Finding: The example prompt "generate Bangla subtitles and add them" is incomplete and vague enough to overlap with normal editing requests, making accidental routing more likely. Because the skill performs remote processing and session setup, ambiguous triggers increase the chance that user content is sent off-platform without sufficiently clear intent.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill encourages users to upload videos and process them through a remote cloud backend, but it does not give a clear warning that files and associated metadata will leave the local environment. This is a material privacy and data-governance issue, especially for personal or sensitive video content where users may not expect third-party transmission.

Natural-Language Policy Violations

Medium

Confidence: 82% confidence
Finding: Hard-coding the session creation request to use `"language":"en"` removes user choice and may cause prompts, UI translations, or processing behavior to occur in a language the user did not select. While not as severe as credential or upload issues, it can lead to user confusion, inaccurate results, and hidden behavior inconsistent with the stated Bangla-focused purpose.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal