PAL Subagent Skill banner
MakiNaruto MakiNaruto

PAL Subagent Skill

Development community

Description

pal-subagent is a one-click skill that installs and registers PAL MCP Server (Claude Code / Codex / others). Once set up, you can delegate simple, well-defined subtasks to an isolated subagent (Codex CLI or Claude Code) via the clink tool — a plan-execute split that cuts main-context token usage.

Installation

This entry records only its repository, not the path inside it, so there is no exact command to give. Open the source below and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

README

PAL Subagent Skill

[English](README.md) | [简体中文](README_zh.md)

Source

Based on the open-source git project [PAL MCP Server](https://github.com/BeehiveInnovations/pal-mcp-server). The install script clones it into `pal-mcp-server/` under this skill directory, with `server.py` as the entry point.

Overview

This project is a **one-click skill for quickly configuring and using PAL MCP Server**:

  • One-click install/uninstall scripts (scripts/) that automatically set up the environment, install dependencies, and register the MCP server
  • Provides the main agent with a "plan-execute" separated subagent delegation capability: via the clink tool of PAL MCP, simple tasks are delegated to an isolated subagent (Codex CLI or Claude Code), saving main-context tokens

Usage

1. One-click Installation

Prerequisites:

Dependency Required Notes
conda (Miniconda/Anaconda) Yes Python environment management
git Yes Clones the PAL MCP Server repo
Codex CLI / Claude Code At least one The actual executor of subagent tasks (Codex must be logged in); MCP registration target
Network Yes First installation needs access to GitHub
# Interactive (press Enter to use the default env name pal-mcp-server)
bash /scripts/pal-subagent-install.sh

# Non-interactive (specify env name)
bash /scripts/pal-subagent-install.sh --env-name pal-mcp-server

The script automatically: clones the repo → writes `.env` (custom API endpoints for local models; kept if it already exists) → creates/reuses a conda env (Python 3.12) → installs dependencies → registers the Claude MCP → backs up and writes the Codex `config.toml`.

2. Manual MCP Setup (without running the script)

First prepare the server (clone the PAL MCP Server repo and install dependencies):

git clone --depth 1 https://github.com/BeehiveInnovations/pal-mcp-server.git
conda create -n pal-mcp-server python=3.12 -y
cond