BotKube Fuse

Botkube Fuse is a terminal-based AI copilot for platform engineers. It connects to GCP, GitHub and Kubernetes to manage secrets, summarize CI/CD runs, investigate pod failures and debug local environments from one command line.

Botkube Fuse is a command-line copilot aimed at platform engineers who spend their days jumping between pipelines, clusters, cloud consoles and developer requests. The vendor calls this load "EverythingOps," and Fuse tries to absorb it by letting an engineer describe a task in plain language and receive guidance grounded in their own infrastructure rather than generic advice. In practice the tool is installed locally through Homebrew or curl, with no login or account needed. It then discovers and connects to Google Cloud Platform, GitHub and Kubernetes, with further integrations described as coming later. From there, an engineer types a prompt in the terminal, or runs the built-in scenarios command to see example tasks and adapt them. The showcased use cases are specific. One covers secret management, where a single command creates, rotates and synchronizes secrets across Google Cloud Secret Manager and GitHub Actions. Another summarizes CI/CD runs by reading GitHub Actions runs and logs, surfacing test failures, likely root causes and flaky tests. A third investigates production incidents such as crashing pods in a GKE cluster alongside IAM permission errors, correlating pod status with permission changes. A fourth handles local environment debugging, checking tool installation, authorization and configuration problems when something like listing buckets or opening an SSH session fails. Fuse comes from the team behind Botkube, the Kubernetes monitoring and chat-ops product that forwards cluster events to messaging platforms. Where Botkube lives inside chat tools, Fuse lives in the terminal and focuses on investigation and routine automation. Compared with general-purpose AI coding assistants, its pitch is narrower: domain knowledge of infrastructure and delivery pipelines, with access to the real systems involved. Teams that run mostly on GCP, GitHub and Kubernetes will find the supported surface a close match, while those on other clouds or CI platforms should check the integration list first. Feedback and support run through a Slack channel and an email address.