skill-management

Route tools through a subagent context using a cached index.

3|Updated Feb 5, 2026
One-click install
npx skills add https://github.com/nano-step/skill-manager --skill skill-management-nano-step
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-management
Source: https://github.com/nano-step/skill-manager/tree/main/skills/skill-management
Command: npx skills add https://github.com/nano-step/skill-manager --skill skill-management-nano-step

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The Skill Manager enables efficient routing and execution of tools by isolating tool definitions in a dedicated subagent context, reducing main agent token usage and enabling scalable orchestration.

Core Features & Use Cases

  • Semantic categorization of tools to support diverse domains (Browser, GitHub, GraphQL, Docs, reasoning)
  • Token-saving isolation by routing tool definitions through a subagent, lowering overall token consumption
  • Fast, deterministic tool discovery using the .opencode/skill-tools.json cache

Quick Start

Configure and run the skill-manager to route tools through a token-saving subagent context for scalable, deterministic tool execution.

Frequently Asked Questions about skill-management

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

FAQPage Schema
How do I reduce token usage when routing and executing multiple tools in an agent workflow?

Token usage during tool routing and execution is reduced by isolating tool definitions in a dedicated subagent context. This approach offloads tool execution overhead from the main agent, minimizing overall token consumption while maintaining scalable orchestration.

What is semantic categorization for tool routing and how does it work?

Semantic categorization for tool routing groups tools into domains like Browser, GitHub, GraphQL, and Docs. This classification enables fast, deterministic tool discovery using a cached index, ensuring safe and precise tool selection across different task types.

How do I set up a tool cache for deterministic tool discovery?

Deterministic tool discovery requires configuring a tool cache at the .opencode/skill-tools.json file path. This cached index allows the routing system to quickly map parameters and select appropriate tools without scanning definitions repeatedly.

Can I route browser, GitHub, and GraphQL tools through a single subagent context?

Yes, browser, GitHub, and GraphQL tools can be routed through a single subagent context. The system applies category-aware routing with strict parameter mapping to ensure deterministic, safe tool selection across these diverse domains.

Why does tool execution require strict parameter mapping?

Strict parameter mapping is required during tool execution to ensure deterministic and safe tool selection. It prevents ambiguous or incorrect parameter passing when routing complex tasks through the subagent context to the target tools.

What are the limitations of using a cached index for tool routing?

The limitation of using a cached index for tool routing is that it requires the .opencode/skill-tools.json cache to be actively maintained. If the cache becomes outdated, tool discovery and category-aware routing may fail to recognize newly added or modified tool definitions.