bootstrap-go-openapi

name	bootstrap-go-openapi
description	Add OpenAPI client code generation to an existing Go project using oapi-codegen. Generates typed Go client and models from an OpenAPI 3.0 spec. Use when the user says "bootstrap openapi", "add openapi client", "generate go client from spec", "oapi-codegen", "openapi codegen", or wants to consume a REST API from Go using a generated client.

Bootstrap Go OpenAPI Client

Add oapi-codegen client generation to an existing Go project. Assumes go.mod already exists.

Prerequisites

Verify before starting:

1. go.mod exists in the project root
2. User has an OpenAPI 3.0 spec (local file or URL)

If the spec is OpenAPI 3.1, warn the user: oapi-codegen has partial 3.1 support. Recommend converting to 3.0 or testing carefully.

Gather Input

Ask the user for:

1. OpenAPI spec path — path to their .yaml/.json spec file (e.g. api/openapi.yaml)
2. Client package name — Go package for generated code (e.g. petstore, acmeapi). Suggest deriving from the API name in the spec.

Detect the module path from go.mod.

Generate

Read references/templates.md for all file templates.

1. Create client package directory

mkdir -p <client_pkg>

2. Create oapi-codegen config

Write <client_pkg>/generate.yaml from the config template. Uses client: true + models: true.

3. Create generate.go

Write <client_pkg>/generate.go with the //go:generate directive pointing at the spec.

The spec path in the directive must be relative to the generate.go file. If the spec is at api/openapi.yaml and generate.go is at petstore/generate.go, the path is ../api/openapi.yaml.

4. Create apiconfig package

Write internal/apiconfig/apiconfig.go from the template. This provides per-environment config storage (base_url + api_key) using Viper's config file.

mkdir -p internal/apiconfig

5. Create configure command

Write cmd/configure.go from the template. Adds three subcommands:

• configure set --env <name> --base-url <url> --api-key <key> — save credentials for an environment
• configure show — list all configured environments (API keys are masked)
• configure use <name> — switch the default environment

Config is stored in the Viper config file (~/.{{APP_NAME}}.yaml) under:

default_env: production
environments:
  production:
    base_url: https://api.example.com
    api_key: sk-xxx

6. Install dependencies

go get github.com/oapi-codegen/runtime@latest
go mod tidy

7. Run code generation

go generate ./...

8. Add justfile recipes

If a justfile exists, append generate and regen recipes from the templates. Check whether a tidy recipe already exists — if so use regen: generate tidy, otherwise inline go mod tidy in the regen recipe. Skip recipes that already exist.

9. Verify

go build ./...

Output

After generation, print:

Added OpenAPI client in <client_pkg>/:
  generate.yaml              — oapi-codegen config (client + models)
  generate.go                — go:generate directive
  client.gen.go              — generated client and models (do not edit)

Added environment config:
  internal/apiconfig/        — per-env config loader (base_url + api_key)
  cmd/configure.go           — configure set/show/use commands

Setup:
  <app> configure set --env production --base-url https://api.example.com --api-key sk-xxx
  <app> configure show
  <app> configure use staging

Regenerate after spec changes:
  go generate ./...

Notes

• Generated *.gen.go files should be committed to version control
• Regenerate with go generate ./... after any spec changes
• The config file (~/.{{APP_NAME}}.yaml) stores API keys — add it to .gitignore (already handled by bootstrap-go-cli)
• For large APIs, consider splitting models and client — see the split config pattern in references/templates.md
• Use apiconfig.Load(env) in your commands to get the active environment's base URL and API key for creating the generated client