danievanzyl/code-skills/tf-remote-state-outputs

Extract Terraform Remote State Outputs

Scan all .tf files in the repo, extract every output block, and generate (or update) a REMOTE_STATE_OUTPUTS.md file documenting what this project exports for remote state consumers.

Instructions

1. Identify the Project

• Read _backend.tf (or any file containing a backend block) to extract:
• S3 bucket (e.g. rimu-tf-state)
• State key (e.g. ritdu-data-governance-eu1-tf-env/terraform.tfstate)
• Project name — derive from the app_name variable default, the state key, or the directory name
• If no backend config exists, use the directory name as the project name and note that backend is not configured.

2. Find All Outputs

• Glob for **/*.tf (exclude .terraform/ directories).
• Grep for ^output\s+" across all matched files to locate every output block.
• For each output found, read the surrounding block to extract:
• Output name (the string in quotes after output)
• Source file (the .tf filename)
• Value expression (the value = ... line)
• Description (the description = ... line if present)
• Sensitive flag if present

3. Generate Descriptions

For each output:

• If a description attribute exists in the output block, use it verbatim.
• Otherwise, infer a concise description from the value expression:
• module.main_vpc.vpc_id → "VPC ID"
• aws_kms_key.s3.arn → "S3 KMS key ARN"
• aws_ecs_cluster.main.arn → "Main ECS cluster ARN"
• Keep descriptions short (< 80 chars), focusing on what the value represents.

4. Write REMOTE_STATE_OUTPUTS.md

Generate the file at the repo root with this structure:

# Remote State Outputs
 
## Outputs Exported by This Project (`<project-name>`)
 
State location: `s3://<bucket>/<key>`
 
| Output | Source | Description |
|--------|--------|-------------|
| `output_name` | `source_file.tf` | Description of the output |

Formatting rules:

• Group outputs by source file — outputs from the same .tf file should be adjacent in the table.
• Sort groups by filename alphabetically.
• Within each group, preserve the order outputs appear in the source file.
• Wrap output names in backticks.
• Wrap source filenames in backticks.
• If any outputs are marked sensitive = true, append " (sensitive)" to the description.

5. Handle Updates

• If REMOTE_STATE_OUTPUTS.md already exists, overwrite it entirely with the freshly generated content. This file is auto-generated and should always reflect current state.

Example Output

# Remote State Outputs
 
## Outputs Exported by This Project (`ritdu-data-governance-eu1-tf-base`)
 
State location: `s3://rimu-tf-state/ritdu-data-governance-tf-base/terraform.tfstate`
 
| Output | Source | Description |
|--------|--------|-------------|
| `switch_role_administrators_arn` | `switch-role-administrators.tf` | ARN of the administrators IAM role for cross-account access |
| `switch_role_developers_arn` | `switch-role-developers.tf` | ARN of the developers IAM role for cross-account access |
| `kms_s3_arn` | `kms.tf` | S3 KMS key ARN |
| `kms_db_arn` | `kms.tf` | RDS KMS key ARN |
| `main_vpc_id` | `vpc.tf` | VPC ID |
| `main_vpc_cidr` | `vpc.tf` | VPC CIDR block |

Rules

• Only document output blocks — ignore variables, locals, data sources, and resources.
• Exclude any files under .terraform/ directory.
• Do NOT fabricate outputs — every entry must correspond to an actual output block in the code.
• If no outputs exist in the repo, write a note saying "This project does not export any outputs."