Rust adopts an LLM policy that bans autonomous agent contributions outright: "fine to review, not to create"

Rust Blog (Inside Rust)·high signalXBlueskyLinkedInCopy link

On August 5 rust-lang/rust published a project-wide LLM policy built on one bright line — LLMs may answer, analyze, distill, refine, check, suggest and review, but not create. Autonomous agent contributions are banned, along with LLM-generated code in public documentation, mechanically LLM-generated responses to review comments, and soundness-critical changes unless the author is a domain expert; disclosure is mandatory for machine translation, trivial changes, bug discovery, and any public LLM-generated text. Reviewers may close non-compliant PRs without explanation, against a backlog of 1,281 open PRs — the first major systems-language project to codify agent contribution rules rather than handle them ad hoc.