- Home
- Skills
- DevOps & Cloud
- Terraform Plan Review
Works with the AI Agents you already use
Terraform Plan Review
Terraform plan review: lists what a plan will destroy, replace or open (public access, wildcard IAM, encryption off).
$5
Terraform Plan Review
Example session with this skill installed
/terraform-plan-review plan.json
- Read your context and instructions
- Compiled the terraform plan review
Illustrative. The plan is invented; the report below is what the script prints for it. Sensitive attributes show only their names.
TERRAFORM PLAN REVIEW (terraform 1.7.5): 4 create, 2 update, 1 replace, 1 delete
1. Will be DESTROYED
aws_s3_bucket.logs (aws_s3_bucket) [holds data]
2. Will be REPLACED (destroy, then create)
aws_db_instance.main (aws_db_instance) [holds data]
replaced: the old one is destroyed and a new one created; data on it is lost unless it is recreated from a backup
forced by a change in: identifier
changed attributes: deletion_protection, identifier, publicly_accessible, storage_encrypted
!! database is publicly accessible
!! encryption is off (storage_encrypted = false)
!! deletion protection is off
3. Security-sensitive creates and updates
aws_security_group_rule.ssh (aws_security_group_rule)
!! open to the internet (0.0.0.0/0) at cidr_blocks[0]
!! SSH port 22 reachable from the internet
aws_iam_role_policy.admin (aws_iam_role_policy)
!! IAM policy allows broad actions (*)
!! IAM policy allows every action on every resource
4. Other updates in place
aws_instance.web (aws_instance)
changed attributes: instance_type, user_data (sensitive)
aws_lambda_function.api (aws_lambda_function)
changed attributes: memory_size, source_code_hash (known after apply)
5. Other creates
2 x aws_cloudwatch_log_group
Check first: 2 resource(s) that hold data are destroyed or replaced: aws_s3_bucket.logs, aws_db_instance.main. Confirm backups and that this is intended.
Limits: this reads the plan as written. It cannot see what the provider will do at apply time, what depends on a resource outside this plan, or values marked unknown until apply. 'Holds data' comes from a built-in list of resource types and can miss custom or newer ones. It never runs terraform and lists changed attributes by name only; a flag may show the risky setting itself.
Next steps
- Confirm backups and intent for everything labelled holds data before anything is applied.
- Work through each flag with the check in the reference file; a clean report is still not approval to apply.
Connects securely to your tools. The creator never sees your data.
About this skill
Read a Terraform plan the way a careful reviewer would, before anyone applies it. Give it the JSON from terraform show -json and it prints what will be destroyed, what will be replaced (and which change forces it), the security-sensitive creates and updates, the other updates and the other creates. Resources that hold data (databases, buckets, volumes and similar) are labelled, and a "check first" line names the ones being destroyed or replaced. It flags a database that is publicly accessible, encryption switched off, deletion protection off, a rule open to 0.0.0.0/0, SSH or database ports reachable from the internet, a public bucket ACL, force_destroy, and IAM policies that allow every action or everything on everything. A reference file says what to check for each flag.
What it reads and writes: only the one plan JSON file you name; it writes nothing. It never runs Terraform, never contacts a cloud provider and never reads your state. No network connection, no environment variables, nothing to install. Changed attributes are listed by name, not value, so secrets in the plan are not copied into the report (a flag can show the risky setting itself, such as storage_encrypted = false).
Honest limits: it reads the plan as written. It cannot see what the provider will do at apply time, what depends on a resource outside this plan, or values marked unknown until apply. The "holds data" label comes from a built-in list of AWS, Google Cloud and Azure resource types and can miss custom or newer ones. The flags go by attribute name, at most six per resource, and the IAM check reads JSON policy documents only. A clean report is not approval to apply. Needs Python 3.9+.
Needs Python 3.9+. Reads the plan as written: it cannot see provider behaviour at apply time, resources outside the plan that depend on these, or values unknown until apply. The holds-data label uses a built-in list of types. A clean report is not approval.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 15 days ago
Requires
Permissions
File Scopes
Reads only the plan JSON file you name; it never runs Terraform, never contacts a cloud, makes no network connection and reads no environment variables. Words such as 0.0.0.0/0, wildcard and encryption appear in the code as patterns it reports.