Amazon recently announced Kiro Crew, an open-source system for running multiple Kiro coding agents across sessions, tools, and tasks. The new workspace lets developers assign asynchronous coding tasks to AI agents, allowing work such as incident investigation, ticket triage, migrations, and PR monitoring to continue without active supervision.
Kiro Crew supports persistent, multi-session development work through shared memory, reusable skills, scheduled jobs, concurrent agents, and purpose-built Apps. Agents can retain project context, run tasks while developers are away, delegate work to subagents, and integrate with external tools through MCP and webhooks.
Developed internally at Amazon as MeshClaw and reportedly adopted internally by more than 39000 developers, Kiro Crew is now available as an open source project and can run locally or on infrastructure developers control. Bolin Chen, senior software engineer at Amazon, Zejiang (Joe) Guo, principal SDE at AWS, and Zezhen Xu, senior software engineer at Amazon, write:
Three of us wanted something simple that wasn’t available internally, a way to kick off a task, walk away, come back to something worth reviewing, and to run several tasks at once instead of babysitting one prompt at a time. We were inspired by the momentum of OpenClaw and various tools with self-learning agents taking over AI work, but we needed something that met our security requirements for internal development work.
Kiro Crew orchestrates agents through the Agent Client Protocol (ACP), providing live visibility into their work. The Activity view shows each agent’s plan, tool calls, approval gates, and results as they happen. Chen, Guo, and Xu add:
Kiro Crew ships with defense in depth from day one, an OS-level sandbox, denied-by-default commands, suspicious-pattern blocking, input validation, sensitive-path blocking, credential redaction, and a signed audit log of every action. Plenty of agents are open source.
Kiro Crew can be customized to fit existing tools and workflows through MCP, Apps, orchestration, and reusable skills. It also supports skills built for other open, standards-based agent platforms without modification.
Kiro Crew's Activity view. Source: Kiro blog.
The community reaction has been mostly positive, with several practitioners highlighting the reported internal adoption. Muhammad Ishaq, CTO at Nextbridge, comments:
The internal adoption numbers say more than the feature list honestly, 39,000 builders and 500 contributors in six months without a mandate means people were actually solving real problems with it, not just trying a new tool because it existed. that kind of organic growth inside a company this size is rare and usually means the tool earned its place before marketing got involved.
While many developers are experimenting with CI/CD migrations, Dependabot triage, scheduled tasks, and long-running work, some have raised concerns about costs, reporting that Crew can consume tokens significantly faster than Kiro CLI. Mathi M, senior solution architect at Verizon, comments:
Having background subagents handle parallel jobs without blocking the main workflow is exact target engineering builders need. Stacking context across sessions instead of starting from scratch will save tons of redundant prompt engineering.
Kiro Crew runs on the Kiro CLI and can use existing .kiro configurations, including steering files, skills, and custom agents. It is released under the Apache 2.0 license and is available for macOS, Linux, and Windows, with integrations for Slack, Telegram, and WeCom.
According to the announcement, the project expects community participation in development and roadmap planning, with Kiro and AWS engineers initially maintaining the project and external maintainers able to join as contributors emerge.