Security audit

Ultimate AI Media Generator

Security checks across malware telemetry and agentic risk

Overview

The skill is a coherent CyberBara media-generation tool, but it includes an authenticated raw API command and persistent credential handling that deserve review before installation.

Review this before installing if you are uncomfortable with a skill storing an API key on disk, uploading local images to CyberBara, saving/opening generated files automatically, or exposing an authenticated raw API command. Use `--no-open` for safer output handling, avoid uploading sensitive reference images unless you understand CyberBara retention and URL access controls, and rotate the API key if it may have been exposed.

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 Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (12)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 95% confidence
Finding: The skill documentation describes capabilities to read environment variables, read and write local files, make network requests, and invoke shell commands, but it does not declare permissions or present a clear trust boundary. That mismatch can mislead users and host systems about the skill's effective authority, increasing the risk of unintended credential access, local data modification, and external data transmission.

Context-Inappropriate Capability

Medium

Confidence: 92% confidence
Finding: The `raw()` method exposes a generic authenticated API caller that can send arbitrary methods, paths, and JSON payloads beyond the skill's advertised media-generation workflows. In a skill context, this broadens the reachable backend attack surface and can enable unintended operations against the same API using the user's API key, defeating least-privilege expectations.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The README instructs users to store an API key in a local config path but does not warn that the key is a secret, should not be shared, and may be readable by other local users, backup tools, or logs depending on environment configuration. In an agent-skill context, encouraging persistent credential storage without minimum handling guidance increases the chance of secret leakage and unauthorized API usage.

Missing User Warnings

Low

Confidence: 74% confidence
Finding: The README advertises commands to retrieve account balance and recent usage records without clarifying that this accesses potentially sensitive billing and activity metadata. While this is expected product functionality, omitting a privacy warning in a skill designed for agent use can surprise users and expose account activity details in shared terminals, transcripts, or logs.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The README encourages users to tell their AI agent a broad natural-language phrase to install the skill, which can overlap with ordinary user requests and make unintentional installation easier in skill-enabled environments. In the context of agent-integrated tooling, overly permissive trigger phrasing increases the chance of social-engineering or accidental invocation of package installation behavior.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The setup instructions state that the API key is stored at ~/.config/cyberbara/api_key but do not clearly warn about the security implications of persistent plaintext credential storage on disk. This can lead users to expose a reusable secret through local compromise, backups, shared accounts, or overly broad file permissions, especially on multi-user or managed systems.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill states that API keys entered via CLI, environment import, or interactive prompt are persisted to ~/.config/cyberbara/api_key, but it does not present this as a security-sensitive storage action or warn about the persistence implications. Persisting credentials locally without a prominent warning can expose secrets to other local users, backups, sync tools, or accidental reuse on shared systems.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The skill documents that generated media is automatically saved locally and opened by default, but it does not clearly warn users that files will be written to disk and launched unless disabled. Automatic saving and opening can create privacy issues, overwrite expectations about local state, and increase exposure to malicious or unsafe content delivered through generated or remote output files.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The documentation explicitly states that uploaded reference images become reusable URLs, but it does not warn users that these URLs may be publicly accessible or long-lived. In a media-generation skill, users may upload sensitive personal, proprietary, or regulated images, so omission of privacy and retention guidance can lead to unintended data exposure.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The code prompts for an API key and immediately saves it to a predictable local file path, but does not clearly disclose to the user that the secret will be persisted on disk. Even though it attempts to set restrictive permissions, silent credential persistence can expose secrets to other local processes, backups, sync tools, or shared environments where users expected one-time interactive entry only.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: When --from-env is used, the code reads a sensitive API key from the environment and writes it to the local cache without an explicit warning or confirmation. This broadens the exposure surface of a credential that may have been intentionally scoped to the current session, causing unexpected long-term persistence on disk.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The client accepts an arbitrary `base_url` and then sends both `Authorization: Bearer` and `x-api-key` headers, plus uploaded file contents, to whatever host is configured. Without enforcing HTTPS, validating trusted hosts, or warning users, this can leak credentials and sensitive media to attacker-controlled or insecure endpoints via misconfiguration or malicious skill configuration.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.