Gateway Direct Execution Skill

AI orchestration request classification differentiates between built-in commands, direct tool invocations, model-driven responses, workflow execution, and agent-required tasks to route user messages to the appropriate execution path.

Standardize execution results by wrapping all direct and model-driven executions in a structured ResultEnvelope, ensuring consistent formatting and deterministic outcomes across the entire processing pipeline.

Manage provider and executor selection by implementing an execution gateway that handles the decision-making process, establishing clear boundaries between request classification, planning, and execution.

Maintain channel affinity throughout the request processing pipeline by utilizing an orchestration runtime that preserves the communication context from initial request classification through to the final ResultEnvelope response.

Workflow compilation is needed when executing pre-compiled workflows, while direct tool invocation handles explicit commands, and model-driven responses manage general queries within the orchestration runtime.

Yes, the request planning approach explicitly distinguishes between built-in commands and agent-required tasks, classifying each request type to ensure the execution gateway routes it to the correct provider.