pi-subagents
Claude Code like Sub-Agents & Workflow Orchestration for Pi — parallel execution, live widget, fleet view, custom agent types, mid-run steering, claude compatible dynamic workflows and more ...
pi-subagents is a Pi extension that brings Claude Code-style autonomous sub-agent orchestration and workflow scripting to Pi. It enables developers to spawn specialized agents that run in isolated sessions—each with its own tools, system prompt, model, and thinking level—executing either in parallel background processes or blocking foreground runs. The extension provides deterministic JavaScript workflow scripts, live UI widgets showing agent activity, FleetView navigation, and Claude Code-compatible APIs. Custom agent types can be defined with YAML frontmatter, and agents support features like git worktree isolation, persistent memory, mid-run steering, and scheduled execution.
Key Features
Use Cases
- 01Delegating parallel code exploration tasks to multiple read-only Explore agents across different modules
- 02Running security audits with custom auditor agents that have specific tool restrictions and system prompts
- 03Orchestrating multi-phase workflows where discovery agents feed findings to verification agents in a pipeline
- 04Isolating refactoring work in git worktrees so experimental changes land on branches without touching the main checkout
- 05Scheduling recurring agents to run weekly commit reviews or periodic codebase scans via cron expressions
- 06Building nested agent hierarchies where a planner delegates to specialist agents with scoped permissions
Related Plugins
View moreandrej-karpathy-skills
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
claude-mem
Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
Understand-Anything
Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
rtk
CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies
pi-subagents — FAQ
What is pi-subagents?+
pi-subagents is a Pi extension that adds Claude Code-style autonomous sub-agent orchestration, enabling you to spawn specialized AI agents with their own tools, models, and system prompts that run in parallel or scripted workflows.
How do I install pi-subagents?+
Run 'pi install npm:@tintinweb/pi-subagents' from the command line. It requires Pi version 0.84.0 or newer.
Which AI clients does pi-subagents work with?+
It works with Pi (pi.dev) as a native extension. Third-party adapters exist for DeepSeek Harness (dsh), though those are maintained independently and not officially tested.
Do I need API keys or prerequisites?+
You need Pi 0.84.0 or newer installed. No additional API keys are required beyond what Pi itself needs for model access. Git is optional but required for worktree isolation features.
Is pi-subagents free to use?+
Yes, the extension is MIT licensed and free to use. However, you still pay for the AI model usage (tokens) consumed by the sub-agents according to your model provider's pricing.
How do I create custom agent types?+
Create .md files in .pi/agents/ (project) or ~/.pi/agent/agents/ (global) with YAML frontmatter specifying name, tools, model, system prompt, and other configuration. You can also use the /agents interactive menu to generate agents via wizard or AI.
How do I install pi-subagents?+
Open the source repository on GitHub and follow its README. pi-subagents is a plugin — MCP Agents Market links you directly to the official repo.
Is pi-subagents free?+
pi-subagents is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.