Skip to main content

atmos toolchain lock

Generate or refresh toolchain.lock.yaml for tools already configured in .tool-versions, without extracting or installing binaries into the toolchain tree. This is the command to reach for when you've turned on toolchain.use_lock_file for an existing project and want to populate the lock file for tools that are already installed — without forcing a full --reinstall of every one of them.

atmos toolchain lock --help

Usage​

Execute the toolchain lock command like this:

atmos toolchain lock

Why a Separate Command​

atmos toolchain install only writes a lock entry as a side effect of actually installing a tool — and it skips tools that are already installed unless you pass --reinstall. That means turning on use_lock_file: true for a project whose tools were installed before the lock file existed leaves those entries missing until every tool is forced to reinstall.

lock downloads and verifies each tool's artifact — the same way install does, to compute a real checksum — but stops short of extracting or placing a binary in the toolchain tree. It only writes/updates the tool's toolchain.lock.yaml entry, so it works the same whether the tool is already installed or not, and never touches an already-installed binary.

Examples​

# Lock every tool in .tool-versions
atmos toolchain lock

# Lock specific tools
atmos toolchain lock terraform
atmos toolchain lock terraform kubectl

# Limit this batch to two concurrent locks
atmos toolchain lock --max-concurrency 2

Arguments​

tool... (optional)

One or more tools to lock. Each tool can be a short name (alias) or full owner/repo format. If omitted, locks every version configured for every tool in .tool-versions.

Examples: terraform, hashicorp/terraform, kubectl

Flags​

--max-concurrency (optional)

Maximum number of tools to lock concurrently. The default is 4. Use a positive integer; values lower than 1 are rejected. Resolved the same way as install's --max-concurrency: --max-concurrency, ATMOS_TOOLCHAIN_MAX_CONCURRENCY, toolchain.max_concurrency, then 4.

Related Commands​