request-optimizer

The Request Optimizer v2.0 serves as the strategic orchestration layer for the Claude Code ecosystem. By intercepting incoming user requests, it acts as a central decision engine that determines the optimal path for task completion. It is designed for developers and project managers who require structured, automated, and model-aware task delegation. The skill evaluates incoming instructions through a rigorous 5-point analysis framework that assesses request specificity, exploration requirements, and potential subtask decomposition.

• Performs automated complexity scoring from 0.0 to 1.0 using weighted factors including scope, depth, ambiguity, tooling requirements, and estimated duration.

• Executes intelligent routing by matching tasks to the appropriate specialized agents (28 available), operational skills (33 available), and MCP server integrations (e.g., obsidian-docs).

• Automates model selection by dynamically choosing between Haiku, Sonnet, or Opus based on the calculated complexity score to balance cost and performance.

• Enforces safety through strict approval gates for high-risk operations such as write-heavy MCP tasks, usage of Opus or LimitlessAgent, and multi-step workflows.

• Prepares for future enterprise-grade scaling through planned integration interfaces for LimitlessAgent environments.

• Input: User request text or project scope. Output: A structured recommendation report including complexity scores, model selection, resource allocation, and a step-by-step execution plan.

• Practical usage: Invoke this tool for non-trivial development tasks or multi-tool workflows to prevent over-allocation of expensive models on simple tasks.

• Constraints: Always requires explicit user approval for destructive actions, heavy exploration, or escalation to high-tier AI models.

• Integrates with external resource registries to ensure routing remains current with the evolving ecosystem of Claude Code agents and command-line tools.