How to use Claude to build a Rust log processor

Kelora began as an attempt to rewrite the author's Python log-processing tool, klp, in Rust. The author barely knew Rust and turned the project into an experiment in how far AI-assisted development could go. The supplied post says the code and extensive test suite were generated entirely by Claude CLI and Codex CLI.

The human supplied feature ideas and discussed them with the agents. The agents then wrote specifications, implementations, tests, documentation, and CI. The author's feedback came from using the program and reviewing its documentation rather than reading its code. The resulting command-line tool processes log files and streams, with built-in functions for parsing, transforming, and analysis, plus features including JWT parsing and pseudonymisation.

The author calls the result functional and useful, while explicitly treating it as an experiment or prototype. Passing more than a thousand automatic tests and checks including clippy, cargo audit, cargo deny, and cargo fuzz did not resolve the absence of human code review. The post suggests trying the supplied example logs. Who it's for: programmers exploring agent-generated implementations in a language they are still learning, while retaining responsibility for evaluating the result.

Prompt
Suggested prompt (written by Agents In Use, not the author's)

Source: Hacker News · dloss