Handling Failures as Values in Agent-Generated Code

Описание и применение

Vow – язык программирования без нулей и исключений, использующий Result‑значения для обработки ошибок, повышая надёжность AI‑агентов.

Задача

AI coding agents often overlook failure paths in languages with nulls and implicit exceptions, leading to runtime crashes.

Результат

In October 2026, Vow achieved 99.5% pass rate on fresh hidden tests versus 99.3% for TypeScript, with comparable performance on hard tasks. Runtime errors are formatted with actionable fixes, improving agent reliability.

Рабочий процесс

• 1. Agent writes function returning Result.
• 2. Agent uses ? operator or match to handle Result.
• 3. If failure, Vow outputs detailed error with fix suggestion.
• 4. Agent can immediately apply fix without human intervention.

Доказательства

AI coding agents such as Claude Code, Codex, and Cursor work in a continuous loop: write, run, read the error, fix.

Источник

Vow has no nulls and no exceptions.

Источник

In tests conducted in October 2026 with Claude Haiku 4.5, code that agents finish passes fresh hidden tests at 99.5% for Vow and 99.3% for TypeScript.

Источник

Every Vow error says what is wrong, where, and the likely fix.

Источник

The command vow --sandbox checks capabilities before anything runs and refuses a program that reaches beyond what you allow.

Источник

Vow compiles to JavaScript. Recursive fib(32) runs in 0.09 s, against 0.14 s for Python.

Источник