climate-generator

Skill: climate-generator

You have access to climate, a CLI that generates production-ready Go command-line clients from OpenAPI specifications and can emit Markdown prompts for agent skills.

When to use this skill

• The user has an OpenAPI 3.x URL or file and wants a CLI quickly.
• The user wants a human-usable API client rather than writing SDK glue code.
• The user wants to turn a generated CLI into a reusable agent skill.
• The user wants to compose multiple microservice specs into one facade CLI.
• The user wants a local OpenAPI simulator (mock server) for testing.
• The user wants the generated CLI to receive or emit webhook callbacks itself.
• The user wants shell completion for climate itself.
• The user wants to list, remove, or upgrade a previously generated CLI.
• The user wants to uninstall climate itself and clean up local climate-managed artifacts.

Core workflow

1. Generate a CLI from the provided spec.
2. If user provides multiple specs, use climate compose instead of climate generate.
3. Capture the resulting cli_name, binary_path, and source_dir.
4. If the user wants agent integration, run climate skill generate <cli-name>.
5. If the user needs sandbox/simulator behavior, run climate mock <openapi_spec>.
6. If the user wants shell completion for climate itself, run climate completion install --shell <shell>.
7. If the user wants the CLI managed on GitHub, run climate publish <cli-name>.
8. Follow the generated instructions from that Markdown prompt.

Commands

Generate a CLI

climate generate [--name <cli-name>] [--out-dir <dir>] [--no-build] [--force] <openapi_spec>

• <openapi_spec> can be a local path or an HTTP(S) URL.
• --no-build generates source only.
• --force overwrites an existing output directory.

Success output is JSON:

{
  "cli_name": "<name>",
  "binary_path": "<absolute path to compiled binary>",
  "source_dir": "<absolute path to generated source>",
  "version": "<API version>",
  "openapi_hash": "<sha256 of the spec>"
}