kirodotdev/KiroCrew

A persistent workspace for development work that self-improves and continues beyond one session.

What it solves

Kiro Crew provides a persistent, self-learning development workspace that continues working even when the user is not actively chatting. It solves the problem of ephemeral agent sessions by running on hardware the user controls, allowing agents to maintain memory, execute long-running tasks unattended, and evolve their skills based on on-going work.

How it works

The system uses a "Gateway" architecture that separates the agent runtime from the user interface. The Gateway manages agent sessions, memory, schedules, and security policies. Users can interact with these agents via a desktop app, web dashboard, CLI, or integrated messaging platforms (like Slack, Discord, and Telegram). Each session is powered by kiro-cli and can spawn isolated subagents for parallel work. The runtime can be hosted locally or on a remote Linux server.

Who it’s for

It is designed for developers who need an AI agent that can handle multi-step tasks, manage recurring jobs, and maintain a long-term project context across multiple sessions.

Highlights

  • Persistent Memory: Sessions and task checkpoints survive restarts, and corrections become durable lessons.
  • Unattended Autonomy: Supports scheduled recurring jobs and reactive work triggered by webhooks or messaging events.
  • Self-Evolving Skills: Synthesizes reusable skills from repeated patterns that can be edited by the user.
  • Multi-Surface Access: Unified state across a desktop app, web dashboard, CLI, and various messaging integrations.
  • Defense in Depth: Includes tool approvals, OS sandboxing, and credential redaction for secure execution.

Related

  • Project
    SII-Holos/synergy

    Synergy is an open‑source, local‑first platform that provides a durable workspace for AI agents. It keeps sessions, files, a built‑in browser, notes, and reusable memory together, and lets you run, schedule, and coordinate agents from a web UI, desktop app, or CLI. The project includes a monorepo built with Bun/TypeScript, supports plugins/skills, and claims benchmark improvements on DeepSWE v1.1.

  • Project
    Team-Commonly/commonly

    An open-source workspace for collaborating with AI agents from any runtime, providing them with persistent identity and shared project memory to eliminate redundant context-setting.

  • Project
    EverMind-AI/Raven

    A self-improving agent harness that allows AI agents to continuously refine their tools, skills, and memory across sessions to evolve into specialized digital workers.

  • Project
    peterfei/ai-agent-team

    A virtual AI development team with specialized roles (PM, Backend, Frontend, etc.) and a persistent memory system for managing complex software projects.

  • Project
    NousResearch/hermes-agent

    A self-improving AI agent with a built-in learning loop that creates and refines skills from experience and maintains persistent user memory across sessions.