fal-ai-media
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to process user-supplied natural language prompts and pass them to external media generation models (images, video, audio). This creates a surface for indirect prompt injection if the user input contains malicious instructions designed to influence the agent's behavior or the output of the generation tools.
- Ingestion points: User prompts provided to the
generate,search, andfindtools as seen inSKILL.md. - Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are used when interpolating user data into tool parameters.
- Capability inventory: The skill utilizes tools for media generation (
generate), model discovery (search,find,models), and file management (upload). - Sanitization: The instructions do not describe any input validation or sanitization, relying instead on the safety filters of the underlying fal.ai and ElevenLabs platforms.
- [EXTERNAL_DOWNLOADS]: The skill documentation instructs the user to configure an MCP server using
npx -y fal-ai-mcp-server. This command fetches and executes the official fal.ai MCP server package from the public npm registry at runtime. - [COMMAND_EXECUTION]: The
~/.claude.jsonconfiguration block specifies the use ofnpxto launch the fal-ai MCP server, involving shell command execution to initialize the skill's capabilities.
Audit Metadata