dot-pi¶
Custom pi agent teams as dotfiles.
What is this?¶
dot-pi is a dotfiles-style repository for managing multiple pi coding agent configurations. Instead of cluttering ~/.pi/ with extensions, agents, and prompts, this repo defines self-contained team directories and standalone agent directories -- each with its own extensions, skills, and session history.
The p command sets PI_CODING_AGENT_DIR to the right directory, and you get a fully isolated pi agent configuration from any working directory.
Quick Start¶
# Clone the repo
git clone git@github.com:PlebeiusGaragicus/dot-pi.git ~/dot-pi
# Source bash_aliases (defines `p`)
echo 'source ~/dot-pi/bash_aliases' >> ~/.zshrc
source ~/dot-pi/bash_aliases
# Set up API keys
cp ~/dot-pi/example.env ~/dot-pi/.env
# Edit .env with your API keys
# Use a team
cd /any/project
p recon "Find all authentication code"
p blog "Write a post about this project's architecture"
Teams¶
Teams are multi-agent configurations with orchestrated delegation (single, parallel, chain). Each team has its own set of specialized subagents, prompt templates, and an orchestrator system prompt. See the Teams section in the sidebar for details on each team.
Standalone Agents¶
Standalone agents are single-purpose configurations with custom extensions instead of subagent orchestration. See Standalone Agents for the concept and individual agent pages.
Creating New Configurations¶
./setup.sh create my-team # new team
./setup.sh create --workspace my-research-team # new workspace team
./setup.sh create-agent my-agent # new standalone agent
All configurations are invokable via p <name> after re-sourcing bash_aliases. Workspace agents (those with a workspace.conf file) launch in a fresh dated directory under workspaces/ and support --list and --resume. See the Usage Guide for details.