The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill utilizes a bash script and Terraform CLI to manage infrastructure. These are expected and legitimate operations for this skill's primary purpose.
[EXTERNAL_DOWNLOADS] (SAFE): The HCL configuration references the 'hashicorp/aws' provider. Per [TRUST-SCOPE-RULE], HashiCorp is a trusted organization, and the use of their official provider registry is considered safe.
[DATA_EXFILTRATION] (SAFE): No evidence of hardcoded credentials or unauthorized data access was found. The S3 backend configuration for state management is a standard and recommended practice.
[PROMPT_INJECTION] (SAFE): No instructions designed to override agent behavior or bypass safety guardrails were identified within the skill body or metadata.
[INDIRECT_PROMPT_INJECTION] (SAFE): The deploy-terraform.sh script possesses an ingestion surface for untrusted data via its command-line arguments.
Ingestion points: The $ENVIRONMENT and $ACTION variables in deploy-terraform.sh are passed as arguments to the script.
Boundary markers: Absent.
Capability inventory: The script can execute terraform init, plan, apply, and destroy commands.
Sanitization: No explicit sanitization or validation of the input arguments is performed. While this creates a vulnerability surface for shell injection, the severity is assessed as SAFE in the context of a boilerplate automation script intended for infrastructure developers.

terraform-infrastructure