agent-crew

Orchestrate persistent multi-agent teams with context memory across sessions.

Updated Apr 16, 2026
One-click install
npx skills add https://github.com/201983290498/agent-crew --skill agent-crew-201983290498
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-crew
Source: https://github.com/201983290498/agent-crew/tree/main/skills/agent-crew
Command: npx skills add https://github.com/201983290498/agent-crew --skill agent-crew-201983290498

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Agent teams often lose context between sessions, forcing rebuilds and re-briefing. Agent Crew provides a framework to create, persist, and awaken multi-agent teams that remember progress and memory across runs.

Core Features & Use Cases

  • Persistent team storage under .claude/teams/<team_name>/ with team_charter, role memories, progress, and workspace.
  • Role-level memory and skills to tailor behavior per role; per-team skills in .claude/teams/<team_name>/<role_name>/skills/ and project-level skills in .claude/skills/.
  • Automated prompt generation and standardized agent configurations via scripts to bootstrap teams and rehydrate sessions.
  • Use cases: long-running research projects, multi-disciplinary collaboration, and iterative experimentation with memory and planable workflows.

Quick Start

Create your first persistent team by loading this skill and running the Plan to configure it.

Frequently Asked Questions about agent-crew

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

FAQPage Schema
How do I maintain agent memory and context across multiple sessions?

Persistent multi-agent memory is maintained by storing team charters, role memories, and progress under the .claude/teams directory, which allows agents to rehydrate context across runs without re-briefing.

How do I orchestrate multi-agent teams for long-running projects?

Multi-agent teams are orchestrated by using gated prompts to enforce collaboration workflows across planning, execution, and review, generating standardized agent configurations to bootstrap teams for iterative experimentation.

Can I assign role-specific skills and knowledge to different agents in a team?

Role-specific skills are assigned by placing them in the .claude/teams/<team_name>/<role_name>/skills/ directory, tailoring agent behavior per role while sharing project-level skills in .claude/skills/.

What is the best way to bootstrap and rehydrate a multi-agent team workflow?

A multi-agent team workflow is bootstrapped by running automated scripts that generate prompts and instantiate standardized agent configurations, rehydrating session context from persistent workspace files.

Does multi-agent orchestration require external dependencies or frameworks?

Multi-agent orchestration requires no external dependencies or frameworks, relying entirely on internal scripts and standardized configurations stored locally within the .claude directories.