The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the processing of untrusted user data through LLM prompts, which establishes a potential surface for indirect prompt injection.
Ingestion points: User-provided variables such as prompt, messages, and input are used within the AI SDK, OpenAI SDK, and Mastra code examples in SKILL.md.
Boundary markers: The provided code snippets do not implement specific delimiters or instructions to ignore embedded commands within the dynamic data.
Capability inventory: The skill enables network operations to interact with external LLM providers including Anthropic, OpenAI, and Google via the Neon AI Gateway.
Sanitization: The integration examples do not demonstrate explicit validation or sanitization of input data prior to model invocation.
[EXTERNAL_DOWNLOADS]: The skill references and downloads configuration and dependencies from the vendor's official resources.
Evidence: Recommends installing the parent skill via npx skills add neondatabase/agent-skills and refers to documentation hosted at neon.com and models.dev.
[COMMAND_EXECUTION]: Provides instructions for local environment setup and infrastructure management using the Neon CLI.
Evidence: Lists commands for provisioning and environment synchronization, including neon deploy, neon config apply, and neon env pull.

neon-ai-gateway