The Agent Skills Directory

[Command Execution] (MEDIUM): The skill instructs the agent to execute terraform init, terraform validate, and terraform plan on dynamically generated configuration files.
Evidence: Found in SKILL.md under 'Step 5: Validate' and the 'Example Workflow' section.
Risk: terraform init is designed to download external modules and provider binaries. If the skill generates configuration pointing to malicious or compromised sources, this results in arbitrary code execution on the host system.
[External Downloads] (LOW): The skill retrieves metadata and code samples via curl from the Terraform Registry and GitHub.
Evidence: Referenced throughout SKILL.md and references/alternative-discovery-workflows.md.
Source: registry.terraform.io, api.github.com, and raw.githubusercontent.com.
Risk: While these are standard infrastructure domains, the specific organization (terraform-ibm-modules) is not included in the pre-defined list of Trusted GitHub Organizations, requiring caution regarding the integrity of fetched content.
[Indirect Prompt Injection] (LOW): The skill processes data from external APIs to influence its code generation and command execution logic.
Ingestion points: registry.terraform.io (module metadata), api.github.com (repository listings), and raw.githubusercontent.com (template code).
Boundary markers: Absent; the skill does not use delimiters or instructions to ignore embedded commands in the fetched data.
Capability inventory: Subprocess execution of the terraform CLI (SKILL.md).
Sanitization: Absent; the skill lacks validation or escaping logic for the content retrieved from remote sources before using it in configuration generation.

terraform-ibm-modules-solution-builder