The Agent Skills Directory

[SAFE]: The skill enforces defensive infrastructure practices across AWS, GCP, and Azure. It explicitly warns against and provides mitigations for high-risk behaviors, such as committing secrets to version control, using unencrypted Terraform state, and deploying resources without proper tagging or security gates.
[EXTERNAL_DOWNLOADS]: The skill references several reputable open-source security and linting tools, including Checkov, Gitleaks, tfsec, and TFLint. These are well-known services in the DevOps ecosystem and are documented neutrally as part of the skill's recommended validation workflow.
[INDIRECT_PROMPT_INJECTION]: The skill has a defined attack surface because it processes external infrastructure requirements to generate code. However, it manages this risk through a rigorous five-phase workflow (ASSESS to HANDOFF) and requires the emission of a CODE_QUALITY_GATE that specifically checks for security risks before completion.
Ingestion points: Processes infrastructure requirements and existing HCL/YAML configuration files.
Boundary markers: Implements a structured lifecycle and quality gate to validate generated outputs.
Capability inventory: Generates cloud resource definitions and local orchestration files (e.g., Docker Compose).
Sanitization: Recommends automated policy-as-code checks (e.g., OPA, Checkov) to ensure all generated code meets security standards.

scaffold