The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided requirements through YAML files or interactive prompts to generate infrastructure code. This process creates a potential attack surface for indirect prompt injection if requirements are sourced from an untrusted third party. However, this functionality is inherent to the skill's primary purpose as a scaffolding tool.
Ingestion points: Requirements gathering phase in SKILL.md.
Boundary markers: No explicit delimiters or instructions are provided to isolate user-supplied requirement data within prompts.
Capability inventory: Infrastructure code generation, execution of terraform apply and kubectl apply, and project validation via scripts/validate_project.sh.
Sanitization: The skill maps requirements to pre-defined templates; no specific sanitization logic for input values is described.
[COMMAND_EXECUTION]: The skill facilitates legitimate infrastructure management via standard CLI tools (Terraform, kubectl, AWS CLI) and a provided local bash script (scripts/validate_project.sh). These operations are documented correctly for their intended DevOps use cases and do not involve unauthorized privilege escalation or malicious activity.
[EXTERNAL_DOWNLOADS]: The skill instructs the agent to retrieve latest versions of various addons and modules from well-known technology services and official repositories, including ArtifactHub, GitHub, and the Terraform Registry. These network operations are documented neutrally and are essential for maintaining the security and currency of the generated environment.

eks-build