The Pulse
Kiro Adds Long-Running Bots to Kiro Crew
Kiro has made customizable, long-running bots generally available in Kiro Crew. Kiro says bots can run scheduled work, coordinate parallel sessions and keep separate memory, tools and permissions.
AI.info Team · Written with AI · fact-checked ·
Kiro describes bots as long-running agents with their own chat and workspace
In a blog post dated November 5, Kiro said that bots are generally available in Kiro Crew. The company describes a bot as a named, long-running agent that users define and customize for a particular part of their work. Kiro says each bot has its own memory and runs its own sessions, so users stop tracking separate agent sessions one by one.
Kiro says a bot has a name, an avatar and one ongoing chat where users can set direction and return for results. It also keeps notes for tracking its work, so users can see decisions and open items. According to Kiro, everything a bot owns lives in its own workspace in Kiro Crew, with a customizable dashboard for seeing where the work stands without opening individual sessions.
Kiro says bots can run parallel sessions and scheduled work
Kiro says a bot can start sessions in parallel, monitor them, redirect anything drifting from the plan and bring results back to one ongoing chat. If a task needs the user's decision, a bot can ask while it keeps working on the parts that do not depend on the answer. Users can also set up schedules in the chat, so a bot can investigate overnight build failures before the user starts the day, or prepare release notes every Friday.
Kiro describes a build-health bot that checks a CI pipeline every 30 minutes, sorts new failures from known flaky tests and only brings back what needs attention. Scheduled activity goes to the bot's work log, while a finding or decision that needs the user returns to chat. The company says users can correct a bot that chases an expected failure, and future runs will apply that distinction.
Kiro says tools and permissions are set per bot
Kiro says each bot gets only the tools and permissions its job needs. A research bot, for example, can search the web and read files but cannot change them; a deployment bot can run deployment scripts and asks for approval before each change. A bot can ask for more tools, but nothing changes until the user approves the request.
Because each bot has its own memory, Kiro says, what a deployment bot learns about a release process does not mix with what a research bot learns. Kiro says users can view, edit or delete anything a bot remembers, adjust its rules, add tools or restrict its permissions. Its examples include an intake bot that can comment on tickets but can't close them, and a code review bot that can leave comments and request changes but can't approve or merge.
Kiro says models, dashboards and custom agents shape the setup
Kiro says users choose the model each bot runs on and can pick a different one for its worker agents. A bot can plan on a stronger model while its worker agents handle routine checks on a faster, cheaper one, according to the post. Kiro says bots can coordinate parallel sessions across worker agents on their own, and are only as proactive as users allow. Related sessions are available for users to review.
Kiro says users can tell a bot how they want to see activities or reports, and the dashboard will show the current state of that work in one place that the bot maintains. Kiro's example shows a GitHub issue bot with a goal to monitor the GitHub issue backlog and report what's trending, alongside a dashboard for the kirodotdev/KiroCrew repository. Kiro says that to start, users open the bots workspace in the Kiro Crew sidebar, add a bot and choose the default or a custom agent as its base; a custom agent supplies the prompt, model, tools and skills.