The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes kubectl and aws CLI commands to gather cluster metadata, workload configurations (Deployments, StatefulSets, etc.), and ECR image scan findings. These operations are performed to facilitate the best practice assessment and are restricted to data collection within the user's configured environment.
[EXTERNAL_DOWNLOADS]: Retrieves current best practices from well-known and trusted services, including the official Kubernetes documentation via context7 and the EKS Best Practices Guide via aws-knowledge-mcp-server.
[INDIRECT_PROMPT_INJECTION]: The skill acts on data retrieved from the EKS cluster (such as pod specifications and events) and documentation from MCP tools, which are external sources.
Ingestion points: Workload configuration JSON from kubectl, image scanning results from aws ecr, and documentation content from MCP search tools.
Boundary markers: Not explicitly defined in the processing logic for cluster data.
Capability inventory: Performs read operations on the cluster, executes documentation searches, and writes results to local markdown files using the Write/file tool.
Sanitization: The skill processes structured data (JSON) and applies version-aware filters, which provides a layer of validation for the ingested cluster information.

eks-workload-best-practice-assessment