Gateway Direct Execution Skill

Classify and plan request execution for an AI orchestration runtime.

22|7|Updated Feb 23, 2026

Search Tags:#orchestration #execution #classification #gateway #request planning #result envelope

Ready to run

npx skills add https://github.com/berrzebb/SoulFlow-Orchestrator --skill gateway-direct-execution-skill

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: Gateway Direct Execution Skill
Source: https://github.com/berrzebb/SoulFlow-Orchestrator/tree/main/.claude/skills/gateway-direct-execution
Command: npx skills add https://github.com/berrzebb/SoulFlow-Orchestrator --skill gateway-direct-execution-skill

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill addresses the complexity of routing and executing diverse requests by establishing clear boundaries between request classification, planning, and execution, ensuring deterministic and standardized outcomes.

Core Features & Use Cases

Request Classification: Differentiates between various request types like builtin, direct_tool, model_direct, workflow_compile, workflow_run, and agent_required.
Execution Gateway: Manages the decision-making process for provider and executor selection.
Standardized Results: Ensures all direct or model-driven executions return a structured ResultEnvelope.
Use Case: When a user sends a message that could be a direct command to a tool, a request to run a pre-compiled workflow, or a general query requiring an agent, this skill intelligently routes it to the most appropriate execution path and formats the response consistently.

Quick Start

Use the Gateway Direct Execution skill to classify and plan the execution of a user's request.

Frequently Asked Questions about Gateway Direct Execution Skill

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How does AI orchestration request classification work for routing user messages?▼

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.

How do I standardize execution results from diverse AI orchestration requests?▼

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.

What is the best way to manage provider and executor selection in an AI runtime?▼

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.

How do I maintain channel affinity throughout an AI request processing pipeline?▼

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.

When do I need workflow compilation versus direct tool invocation in request planning?▼

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.

Does this request planning approach support distinguishing between built-in commands and agent-required tasks?▼

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.