The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill includes explicit security guidelines to prevent credential exposure. It instructs the agent to confirm that authentication is provided through secure variables rather than checked-in literals and specifically warns against adding tokens (e.g., KONNECT_TOKEN, KONNECT_SPAT) to .tf files.
[COMMAND_EXECUTION]: The skill utilizes standard Terraform CLI commands such as terraform plan, terraform apply, terraform validate, and terraform providers schema. These are legitimate and necessary for the skill's stated purpose of infrastructure management.
[SAFE]: The instruction to 'retry outside the sandbox with approval' if provider startup fails is documented as a troubleshooting step for environment-specific interference. Because it explicitly requires user approval and targets standard development tools, it does not constitute a malicious sandbox bypass.
[SAFE]: All referenced external resources, such as the kong/konnect and kong/konnect-beta Terraform providers, originate from the official vendor (Kong) and are appropriate for the skill's functionality.

terraform-konnect