skills.rest
ai-pivot
ai-pivotOFFICIAL
Explore all skills by this author →
GitHub

mod-tools

Provides unified registry and lifecycle management for MCP-connected tools.

12|5|Updated Apr 26, 2026
Search Tags:#MCP integration#LLM context#tool execution#tool registry#agent tools#Feishu OAuth
One-click install
npx skills add https://github.com/ai-pivot/xbot --skill mod-tools
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: mod-tools
Source: https://github.com/ai-pivot/xbot/tree/main/.xbot.example/skills/mod-tools
Command: npx skills add https://github.com/ai-pivot/xbot --skill mod-tools

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Tools make an AI agent useful by letting it execute real actions (like calling MCP services, managing tool context, and returning structured results) instead of only generating text.

Core Features & Use Cases

  • Tool registry and execution contract: Defines a unified Tool interface (name, description, parameters, Execute) so the agent can discover and run tools consistently.
  • Standardized results for LLM planning: Produces ToolResult with Summary, Detail, Tips, and WaitingUser to guide the next agent step and UI rendering.
  • MCP-aware integrations (including Feishu): Supports MCP management, session-level MCP loading/unloading, and Feishu MCP wrapping with OAuth-based token handling and enterprise-domain resolution.

Quick Start

Ask your agent to use the Tools module to perform a task via MCP or Feishu integration, and ensure your requested tool is supported by the registered Tool interface.

Frequently Asked Questions about mod-tools

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

FAQPage Schema
How do I execute MCP-managed remote tools from an AI agent?▼

To execute MCP-managed remote tools from an AI agent, you need a consistent Tool interface with parameter schemas and deterministic execution pathways. This allows the agent to discover and run registered tools reliably instead of just generating text.

What is the best way to standardize LLM context for tool execution results?▼

Standardizing LLM context for tool execution requires producing a structured ToolResult. This result format includes Summary, Detail, Tips, and WaitingUser fields to guide the agent's next planning step and ensure correct UI rendering.

Does Feishu MCP integration support OAuth-based token handling?▼

Feishu MCP integration supports OAuth-based token handling alongside enterprise-domain resolution. It wraps Feishu services within the tool registry to allow agents to securely execute actions using tenant-derived domains.

Can I load and unload MCP tools dynamically during an agent session?▼

You can load and unload MCP tools dynamically during an active agent session. This session-level MCP management capability allows the agent workflow to adapt its available external capabilities on demand.

How do agent tools ensure safe URL construction for external API calls?▼

Agent tools ensure safe URL construction by using client helpers that derive domains from the tenant context. This deterministic pathway prevents malformed requests when executing registered external capabilities.

Why do I need a tool registry for AI agent workflows?▼

A tool registry is needed for AI agent workflows to provide a unified execution contract. By defining a consistent interface with names, descriptions, and parameters, the agent can reliably discover and run external actions.

Related Skills

Explore More Software Engineering Skills→
RedHatProductSecurityRedHatProductSecurity
47

dynamic-client-registration

Automate OAuth 2.0 client registration requests and validate client metadata.

Official
Intermediate
maslennikov-igmaslennikov-ig
243

parse-error-logs

Parse npm, pnpm, TypeScript, Jest, and build logs into structured JSON reports.

Community
Intermediate
AradotsoAradotso
7

mcp-code-execution-mode

Proxy MCP tool calls through a sandboxed run_python bridge.

Official
Advanced
iani-kuliiani-kuli

native-mcp

Integrates external MCP servers via stdio or HTTP into the agent's unified tool registry.

Community
Advanced