ClawHub
Back to skill

Security audit

Fal.ai API

Security checks for vulnerabilities and agentic risk

Overview

This is a straightforward fal.ai media-generation client, but users should know prompts and media URLs are sent to fal.ai and may use paid API credits.

Install only if you are comfortable sending prompts, image URLs, audio URLs, and related generation parameters to fal.ai using your FAL_KEY. Use a dedicated or limited API key where possible, avoid secrets or regulated personal data in prompts and URLs, and remember that API calls may incur fal.ai charges.

SkillSpector

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
86% confidence
Finding
The skill metadata declares only an environment variable requirement, but the documented usage and capabilities clearly imply outbound network access and shell/CLI execution. This mismatch can weaken policy enforcement and user understanding, causing the skill to operate with broader effective capabilities than its declared permission model suggests.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The README explicitly instructs users to send prompts and generation requests to fal.ai's external API, but it does not disclose that user inputs and possibly referenced media will leave the local environment and be processed by a third party. This can lead to unintentional exposure of sensitive prompts, proprietary content, or regulated data, especially because the skill is framed as a simple media-generation utility without privacy caveats.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill sends user prompts and potentially media inputs to fal.ai for remote processing, but the description does not clearly warn users that their content leaves the local environment. This creates a privacy and data-handling risk, especially if users submit sensitive prompts, images, audio, or other confidential material under the assumption of local-only processing.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The transcription method sends user-provided audio URLs to the external fal.ai service without any in-code notice, consent flow, or guardrails around sensitive content. In a skill handling media, this is contextually plausible functionality, but it still creates a privacy risk because users may unintentionally submit confidential or regulated audio to a third party.

Static analysis

No suspicious patterns detected.