Создание AI-инструмента на Rust с инфраструктурой

★ 7.0 · ai-tooling

create-ai-tool is a Claude Code skill that guides developers through building a new AI tool end-to-end — Rust implementation, toolset wiring, infrastructure setup, JSON schema generation, and frontend UI. The 11-step workflow covers writing a Rust struct deriving `JsonSchema` and implementing `AsyncTool<Context>`, placing it in the correct domain crate under `crates/<crate>/src/inbound/toolset/`, and registering it via `AsyncToolSet::new().add_tool()`. Context dependencies are wired through `FromRef<ToolServiceContext>`, new environment variables and AWS resources are declared in the Pulumi configuration at `infra/packages/shared/src/ai_tools.ts`, and running `cargo fmt`, `cargo clippy`, `bun gen-tools`, and `bun check` validates the full chain. Generated schemas are transpiled into TypeScript for `apps/web`, and the skill then walks through creating a SolidJS tool renderer using the expandable `BaseTool` pattern. It targets teams building Rust-based AI agents on a Rust + TypeScript/Bun + AWS stack who need to extend the toolset without touching the `ai_toolset` framework crate itself.