skills.rest
project-ax
project-axOFFICIAL
Explore all skills by this author →
GitHub

runners

Document and manage agent runner implementations for pi-session and claude-code.

5|2|Updated Feb 8, 2026
Search Tags:#implementation#tooling#agent#LLM#SDK#runner
One-click install
npx skills add https://github.com/project-ax/ax --skill runners-project-ax
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: runners
Source: https://github.com/project-ax/ax/tree/main/.claude/skills/ax/runners
Command: npx skills add https://github.com/project-ax/ax --skill runners-project-ax

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill provides the necessary tools and documentation to understand, modify, and extend the core agent runner implementations within the AX project, ensuring seamless AI agent execution.

Core Features & Use Cases

  • Runner Configuration: Understand how different runners (pi-session, claude-code) are dispatched and configured.
  • LLM Transport: Learn about the mechanisms for LLM communication, including proxy sockets and IPC fallbacks.
  • Tool Management: Discover how tools are registered, filtered, and exposed to different agent SDKs (MCP, pi-coding-agent).
  • Use Case: When developing a new agent type that requires custom LLM interaction or tool integration, consult this Skill to understand the existing runner patterns and how to adapt them.

Quick Start

Use the runners skill to understand the pi-session runner implementation.

Frequently Asked Questions about runners

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

FAQPage Schema
How do I configure an autonomous AI agent runner for custom LLM communication?▼

Agent runner implementations are configured by selecting LLM transport mechanisms like proxy sockets and IPC fallbacks, ensuring seamless communication between the autonomous agent and the language model during execution.

What is the best way to register and expose tools to an agent SDK using MCP?▼

Tool registration for agent SDKs involves filtering and managing tools via MCP and pi-coding-agent SDKs, dictating how specific capabilities are exposed and made available to different autonomous agent types during execution.

How do pi-session and claude-code runners handle stream processing?▼

Pi-session and claude-code runners handle stream processing by managing LLM transport selection and tool registration via MCP and pi-coding-agent SDKs, defining how autonomous AI agents dispatch tasks and process execution streams.

Can I modify an existing agent execution environment to support a new agent type?▼

Modifying an existing agent execution environment to support a new agent type requires adapting existing runner patterns, consulting implementation documentation to integrate custom LLM interactions and tool registration for the new agent.

Does the agent runner implementation support IPC fallbacks for LLM transport?▼

Agent runner implementations support IPC fallbacks for LLM transport, providing a reliable communication channel that activates when primary proxy socket connections encounter issues during autonomous AI agent execution.

Related Skills

Explore More Software Engineering Skills→
agentmattersagentmatters
3

mcp-server-patterns

Automates Node/TypeScript MCP server pattern development, maintenance, and debugging including registration and transport setup.

Official
Advanced
0xHoneyJar0xHoneyJar
5

coding-agent

Execute AI coding agents via bash commands with PTY support.

Official
Intermediate
zengbaochengzengbaocheng

native-mcp

Connect to MCP servers and register their tools in Hermes Agent.

Community
Intermediate
kamalbuildskamalbuilds

agentmarket-autonomous

Enables autonomous AI agents to self-register, monitor, execute, and collect payments on AgentMarket.

Community
Advanced