The Agent Skills Directory

Information Retrieval: The skill fetches technical guidance from official cloud provider documentation and well-known infrastructure registries. This grounding process is a standard practice to ensure accuracy in solution design. Since these sources are reputable, this activity is considered safe.
Deployment Artifact Generation: The skill assists in creating Terraform configurations and verification scripts (using tools like gcloud and curl). These items are generated for the user to review and apply, which is a routine function for an assistant focused on technical implementation.
External Content Processing: The skill processes data from external URLs to generate architecture guides. \n
Ingestion points: Documentation is retrieved in SKILL.md (Phases 2 and 3). \n
Boundary markers: No explicit delimiters are used for external content. \n
Capability inventory: The skill generates design guides, diagrams, and shell scripts. \n
Sanitization: Content is processed without specific filtering. \n
Evaluation: Given that the external data comes from official, highly trusted sources, the risk of this ingestion pattern is low.

google-cloud-solution-agentic-ai-borderless-data-lakehouse