agentsclimarketplace

Terraform bump provider

Skill tony/ai-workflow-plugins/.agents/skills/terraform-bump-provider

Move a Terraform or OpenTofu provider to a new version across every module that declares it, then refresh each root module's lock file without narrowing its platform coverageFrom its SKILL.md

Install
npx -y skills add tony/ai-workflow-plugins --skill terraform-bump-provider

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

7.1 KB, ~1.7k tokens by cl100k_base, as published. Nobody here has run it

Bump a Terraform provider

Raise a provider's version across every module that declares it, in one commit, then refresh the lock file of every affected root module.

Use the terraform-bumping-terraform skill for the phase structure. It reads the same three references this command does, so the two cannot drift: references/layout-discovery.md, references/pin-sites.md, and references/lock-and-init.md.

To move the CLI itself, use the terraform-bump-terraform skill. To refresh locks within the constraints already written, use the terraform-refresh-lock skill.

User arguments: $ARGUMENTS

Context

Repository — run this command and read the output:

git remote get-url origin 2>/dev/null || echo "(not a git repository)"

Directories holding a committed lock file or a backend block — run this command and read the output:

{ git ls-files -- '*.terraform.lock.hcl' 2>/dev/null | sed -E 's|\.terraform\.lock\.hcl$||; s|/$||; s|^$|.|'; git grep -lE '^[[:space:]]*(backend[[:space:]]+"|cloud[[:space:]]*\{)' -- '*.tf' '*.tofu' 2>/dev/null | xargs -r -n1 dirname; } 2>/dev/null | sort -u | grep . || echo "(none — no root module found)"

Source addresses declared in tracked configuration, provider and module alike — run this command and read the output:

git grep -hE '^[[:space:]]*source[[:space:]]*=' -- '*.tf' '*.tofu' 2>/dev/null | sed -E 's/.*"([^"]*)".*/\1/' | grep -vE '^\.{1,2}/' | sort | uniq -c | sort -rn | head -30 | grep . || echo "(none)"

Version constraints declared — run this command and read the output:

git grep -hE '^[[:space:]]*version[[:space:]]*=[[:space:]]*"' -- '*.tf' '*.tofu' 2>/dev/null | sed -E 's/^[[:space:]]*//' | sort | uniq -c | sort -rn | head -30 | grep . || echo "(none)"

Tooling available — run this command and read the output:

{ command -v terraform >/dev/null 2>&1 && terraform -version | head -1; command -v tofu >/dev/null 2>&1 && tofu -version | head -1; command -v terragrunt >/dev/null 2>&1 && terragrunt --version | head -1; } 2>/dev/null | grep . || echo "(no terraform, tofu, or terragrunt on PATH)"

Procedure

Follow the skill's phases. This command supplies the target and the scope.

1. Resolve which provider

The argument may be a full source address or a bare name. Resolve it against the addresses the repository actually declares, inside required_providers blocks — the context above cannot tell a provider source from a module source, and only the block it sits in can. If more than one declared address matches, ask which.

A name that matches nothing declared is a finding, not a typo to correct silently: say so and stop.

2. Discover the layout and settle the ambiguities

Run the skill's discovery phase. --root-module limits the run to the named directories; without it, discover them and confirm the set when there is more than one.

3. Resolve the target version

If the user named a version, confirm it is published. Otherwise take the latest stable release. Either way, record the tag and the source repository the registry reports, and derive the changelog link from that source rather than from a table of providers somebody remembered.

4. Inventory the sites and predict the edits

Every module declaring this provider, root and child, per the pin-sites reference. Decide per site whether it needs editing at all: a constraint that already admits the target does not, and rewriting it anyway narrows what the module accepts.

5. Present the plan and wait

Show the sites to be edited with the operator each keeps, the sites left alone with the reason, the lock files to be refreshed with how many platforms each currently tracks, and any drift found. With --audit-only, stop here — nothing is written.

6. Edit and commit

All constraint edits in one commit. Then, unless --no-lock, refresh each affected root module's lock and commit those separately, checking each lock's per-provider platform count against what it was. With --no-commit, leave everything in the working tree and say what would have been committed.

7. Verify

Read the selected version back out of each lock file and compare it to the target. Run the repository's own gates. Report root modules that did not move, and why.

Rules

  • No version string is written before the release is confirmed to exist.
  • Every declaration of the provider moves in the same commit, or none does.
  • Constraint edits and lock files are separate commits.
  • Each site keeps its existing operator; a bump never normalises operators repository-wide.
  • A lock file never comes back tracking fewer platforms than it had.
  • Report drift and unrelated breakage; fix neither as a side effect.
  • Untracked files are never edited, and .terraform/ is never searched.

Output

Open with a one-line hero (✓ <provider> <old> -> <new> across N root modules or ⚠ Audit only: N sites behind), then exactly these sections:

  1. ## Target — the provider's resolved source address, the target version, its publication date, and the changelog link derived from the registry.
  2. ## Layout — root modules and child modules discovered, which tool drives the repository, and any classification the user settled.
  3. ## Sites — every declaration found, whether it was edited or deliberately left alone, and the operator each kept.
  4. ## Locks — per root module: the version selected before and after, and the platform count tracked before and after.
  5. ## Verification — the gates run and their real results, with any pre-existing failure named as pre-existing.
  6. ## Drift — pin sites that disagreed with each other before the bump, root modules that did not move and why, and modules whose constraints already admitted the target.

End with an ask-user-choice panel offering next steps — for example: converge the pin sites that disagree, refresh the lock files this run left alone, restore multi-platform coverage to a lock that only ever tracked one, or stop here. Skip the panel only in plan mode.

Portability notes

  • ask-user-choice — present the listed options and wait for the user to pick one. Hosts with a structured multiple-choice tool (Claude Code's AskUserQuestion) should use it; otherwise print a numbered list and wait for a numbered reply. Never proceed on an assumed answer.
  • $ARGUMENTS — the text the user passed when invoking this skill. If your host does not substitute it, read it as the user's request in the current turn, and ask when there is none.
  • Bundled files — every relative path in this skill points at a file shipped inside this skill directory. Read them from here, not from the host's plugin tree.

What ships with it: 3 files

20.4 KB alongside SKILL.md

Gives 0 of the 12 instructions most containers cloud skills give in ~1.7k tokens

Counted across 607 of the 657 authors here whose files we hold, read 2026-08-07

  • Run containers as a non-root userin 66 of 607, across 46 files
  • Use multi-stage buildsin 53 of 607, across 44 files
  • Use Promise.all for independent operationsin 47 of 607, across 13 files
  • Import directly instead of barrel filesin 46 of 607, across 12 files
  • Use ternary instead of AND for conditionalsin 45 of 607, across 12 files
  • Use Set or Map for O(1) lookupsin 42 of 607, across 10 files
  • Create a .dockerignore filein 41 of 607, across 31 files
  • Read individual rule files for detailsin 39 of 607, across 9 files
  • Copy dependency files before source codein 36 of 607, across 23 files
  • Authenticate server actions like API routesin 35 of 607, across 7 files
  • Use next/dynamic for heavy componentsin 34 of 607, across 9 files
  • Use React.cache for per-request deduplicationin 34 of 607, across 10 files

Said here and by no other author read

  • resolve the provider source address
  • confirm the target version is published
  • inventory every site declaring the provider
  • present the plan and wait for approval
  • move every declaration in one commit
  • refresh each affected root module lock separately

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 326,852. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.