Function render_json_schema 

pub fn render_json_schema(command: &Command) -> Result<String, Error>

Generate a JSON Schema (draft-07) describing the inputs for a command.

The schema object is suitable for use in agent tool definitions (e.g. OpenAI function calling, Anthropic tool use, MCP tool input schemas).

Arguments appear as required string properties (with "required" if marked so). Flags with crate::model::Flag::takes_value become string properties; boolean flags become boolean properties.

Examples

let cmd = Command::builder("deploy")
    .summary("Deploy a service")
    .argument(Argument::builder("env").required().description("Target environment").build().unwrap())
    .flag(Flag::builder("dry-run").description("Simulate only").build().unwrap())
    .flag(Flag::builder("strategy")
        .takes_value()
        .choices(["rolling", "blue-green"])
        .description("Rollout strategy")
        .build().unwrap())
    .build().unwrap();

let schema = render_json_schema(&cmd).unwrap();
let v: serde_json::Value = serde_json::from_str(&schema).unwrap();
assert_eq!(v["title"], "deploy");
assert_eq!(v["properties"]["env"]["type"], "string");
assert_eq!(v["properties"]["dry-run"]["type"], "boolean");
let strats = v["properties"]["strategy"]["enum"].as_array().unwrap();
assert_eq!(strats.len(), 2);