The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill describes patterns for ingesting data from external cloud provider APIs, remote state backends, and external modules, which represents a potential attack surface for indirect prompt injection if those sources are compromised.
Ingestion points: Data sources (e.g., aws_ami, aws_vpc in examples/core.md), terraform_remote_state (in examples/state.md), and http data sources (in examples/patterns.md).
Boundary markers: While the skill does not explicitly define delimiters for external data ingestion, it strongly encourages the use of validation blocks for all user-supplied variables.
Capability inventory: The skill enables infrastructure management through terraform apply and network health checks via http data sources.
Sanitization: The skill features advanced validation techniques, including precondition and postcondition blocks, to ensure data integrity.

infra-iac-terraform