<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> Pi-Messenger extends the pi coding agent ecosystem, enabling multi-agent communication and coordination across terminals. It acts as a decentralized chat room for AI agents, facilitating discovery, status sharing, task claiming, file reservation, and messaging without a central server. This enhances automated task orchestration and collaborative development for pi agent users.

How It Works

This pi extension hooks into agent lifecycle events, using file-based coordination stored in ~/.pi/agent/messenger/ and project data under .pi/messenger/. It tracks agent activity via pi tool hooks, enabling "Living Presence" (status, usage), an "Activity Feed," and agent discovery. File reservations block write operations, and idle agents are flagged "stuck." The "Crew" feature orchestrates LLM agents (planner, worker, reviewer) as pi --mode json subprocesses, using JSONL streaming for real-time progress and autonomous wave execution based on task dependency graphs. Coordination is file-based, eliminating daemon requirements.

Quick Start & Requirements

• Installation: pi install npm:pi-messenger
• Prerequisites: Requires the pi coding agent environment. LLM API access is necessary for "Crew" task orchestration.
• Resource Footprint: Crew's parallel LLM agent execution can lead to significant token consumption.
• Links: Official documentation and examples are embedded within the README.

Highlighted Details

• Living Presence: Real-time status, tool/token usage, and auto-generated messages for agents and human participants.
• Crew Task Orchestration: Automates workflows by transforming PRDs into task dependency graphs, executed in parallel waves by planner, worker, and reviewer agents. Supports autonomous execution.
• File Reservations: Agents claim files/directories, preventing conflicts and ensuring orderly access, with automatic release on exit.
• Configurable LLM Models: Allows fine-grained control over LLM providers and models for different agent roles to manage costs.

Maintenance & Community

The provided README does not contain explicit information regarding maintainers, sponsorships, community channels (e.g., Discord, Slack), or a public roadmap.

Licensing & Compatibility

• License: MIT.
• Compatibility: The permissive MIT license generally permits commercial use and integration within closed-source projects without significant restrictions.

Limitations & Caveats

The project is an extension of the pi agent ecosystem, limiting standalone applicability. The "Crew" functionality, while powerful for automation, carries a risk of high LLM operational costs due to parallel agent execution. The README lacks specific performance benchmarks or details on community support infrastructure.