skills.rest
panamini
panaminiCOMMUNITY
Explore all skills by this author →
GitHub

pi-messenger-crew

Orchestrate PRD-to-execution workflows across multiple AI agents.

Updated May 4, 2026
Search Tags:#task management#agent coordination#multi-agent orchestration#autonomous workflows#dependency scheduling#crew planning#PRD-driven execution
One-click install
npx skills add https://github.com/panamini/pi-tooling --skill pi-messenger-crew
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: pi-messenger-crew
Source: https://github.com/panamini/pi-tooling/tree/main/extensions/pi-messenger/skills/pi-messenger-crew
Command: npx skills add https://github.com/panamini/pi-tooling --skill pi-messenger-crew

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It solves the problem of coordinating multiple AI agents to convert a PRD or steering prompt into a consistent, dependency-aware plan and then execute tasks reliably.

Core Features & Use Cases

  • Multi-agent Crew Orchestration: Enables orchestrator-driven coordination for planning, task execution, and review using a Crew workflow.
  • PRD-driven Planning & Re-planning: Plans from a PRD file, an inline prompt, or a steering prompt injected into planning notes, with safeguards against revising active in-progress tasks.
  • Dependency-aware Task Lifecycle: Supports creating, starting, progressing, completing, blocking/unblocking, resetting (optionally cascading), and revising tasks and subtrees.
  • Execution Control for Autonomous Work: Runs single-wave or autonomous continuous work with configurable concurrency and model overrides.
  • Stateful Coordination & Persistence: Stores planning, tasks, blocks, configuration, and an activity feed under .pi/messenger/crew/ and .pi/messenger/feed.jsonl for traceability and recovery.

Quick Start

Use the pi_messenger crew skill to plan and start autonomous agent work by running pi_messenger({ action: "plan" }) in the project context that contains your PRD or where you want the planning prompt to scan the codebase.

Frequently Asked Questions about pi-messenger-crew

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

FAQPage Schema
How do I coordinate multiple AI agents for PRD-driven execution?▼

Multi-agent PRD-driven execution is orchestrated by planning a dependency-aware task workflow, then running single-wave or autonomous continuous execution with configurable concurrency and model overrides.

Can I use a steering prompt instead of a PRD file for multi-agent planning?▼

Yes, multi-agent planning accepts a PRD file, an inline prompt, or a steering prompt injected into planning notes to scan the codebase and generate a dependency-aware execution plan.

How does dependency scheduling work for autonomous multi-agent workflows?▼

Dependency scheduling in autonomous workflows manages task lifecycles by creating, starting, blocking, unblocking, and completing tasks, with optional cascading resets for revising task subtrees.

What's the best way to prevent re-planning from disrupting active task work?▼

To protect active task work from re-planning disruptions, the orchestration system includes safeguards against revising tasks that are currently in-progress, ensuring dependency-driven execution continues reliably.

Where are crew artifacts and task states persisted for multi-agent coordination?▼

Crew artifacts, task states, blocks, configuration, and an activity feed are persisted under the `.pi/messenger/crew/` directory and `.pi/messenger/feed.jsonl` for traceability and recovery.

Do I need a YAML frontmatter entry point to use multi-agent crew orchestration?▼

Yes, multi-agent crew orchestration requires a YAML-frontmatter-defined entry point to expose command-scoped actions for planning, task state transitions, and revision workflows.

Related Skills

Explore More Software Engineering Skills→
polikeijipolikeiji

constitution

Create structured project constitutions for coding and documentation standards.

Community
Advanced
sequenziasequenzia
42

execute-tasks

Execute Claude Code Tasks in dependency-aware waves with parallel agents and retries.

Community
Advanced
stepandelstepandel
4

linear-ticket-prep

Prepare Linear tickets for coding agents by validating sizing and generating prompts.

Community
Advanced
AO-Cyber-SystemsAO-Cyber-Systems

df:execute-objective

Execute objective jobs with wave-based parallelism and dependency-aware scheduling.

Official
Advanced