OpenTofu 1.13 starts building linting and policy checks into the IaC workflow

OpenTofu, the open-source, Linux Foundation-hosted Terraform fork, has started building linting directly into the tool. In a September 3 post the maintainers set out their thinking, and the 1.13.0-beta1 release (August 27) ships the first experimental slice: a -lint=all flag on tofu validate, plan and apply that runs four built-in rules — flagging untyped input variables, unused variables and locals, and the old count = cond ? 1 : 0 idiom where the newer enabled meta-argument would do. The more interesting part is the design direction: because OpenTofu's plan phase already evaluates expressions and queries providers, the team argues pre-plan linting and post-plan policy checks are one problem, and wants a single ruleset re-checked at validate, plan and apply time, with rules written in HCL-like syntax, able to pull external data through provider plugins, configurable as blocking or warning-only, and shareable like modules. That pitches it squarely at the ground now covered by tflint, OPA and Sentinel-style policy tooling, and it is the most-voted feature request in the project's tracker. The same beta also adds experimental symbol libraries (reusable functions and types), official Windows on ARM64 builds, drops the WinRM provisioner connection, and is the last series with 32-bit builds. Why it matters for teams: if you run OpenTofu in CI, a native, versioned lint/policy layer means one fewer tool to pin and one place to encode house rules — worth trialling on a non-production stack now so feedback shapes the RFC.

This is a summary by our content curator. Read the original at OpenTofu: https://opentofu.org/blog/plans-for-linting/.