vfarcic/dot-ai

DevOps AI Toolkit – AI‑powered platform engineering for Kubernetes

What it is – A Node‑JS library (published as @vfarcic/dot-ai) that adds large‑language‑model (LLM) intelligence to DevOps workflows. It ships a set of services and a CLI that can be plugged into a Kubernetes‑based platform (the MCP – Managed Cloud Platform) to let operators interact with clusters, get deployment advice, and automate remediation using natural‑language prompts.

Core capabilities

• Natural‑language cluster queries – ask questions like “Which pods are using more than 80 % CPU?” and receive structured answers.
• Deployment recommendations – the engine analyses manifests and suggests improvements (resource limits, best‑practice settings, etc.).
• Issue remediation & root‑cause analysis – when a failure is detected, the AI proposes corrective actions and explains why.
• Policy & pattern management – store organisational policies as reusable prompts that can be applied automatically to new repositories.
• Semantic documentation search – index internal docs and retrieve relevant sections via plain‑text queries.
• Repository bootstrap – creates new repos pre‑populated with governance files (e.g., CODE‑OF‑CONDUCT, SECURITY.md).
• Shared prompt libraries – a central catalogue of prompts ensures consistent AI‑driven workflows across teams.
• Untrusted‑content boundary – isolates raw cluster output from the AI‑generated instructions, reducing the risk of executing malicious suggestions.

How you get it

• Full stack install – a one‑click script (referenced in the Stack Installation Guide) provisions all required components (API server, CLI, prompt store, etc.) on a Kubernetes cluster.
• Individual components – the Deployment Guide explains how to install only the pieces you need (e.g., just the CLI).
• The package is published on npm (@vfarcic/dot-ai) and can be added to any Node project.

Typical workflow

1. Connect the CLI or MCP UI to a Kubernetes cluster.
2. Ask a natural‑language question or request a task (e.g., “Create a Helm chart for a Redis cache with high‑availability.”).
3. The AI Engine (built on the Model Context Protocol and Vercel AI SDK) calls an LLM, processes the response, and returns actionable output.
4. Optionally apply the generated manifests or remediation steps directly to the cluster.

Who benefits

• Platform engineers who want faster insight into cluster state without writing kubectl queries.
• SRE teams needing quick, AI‑assisted root‑cause analysis for incidents.
• Development teams that want automated, policy‑compliant repo scaffolding.
• Organizations looking to embed consistent AI‑driven best practices across many Kubernetes environments.

Project health

• Beta status (orange badge) – core features are usable but may still evolve.
• Active npm releases, GitHub stars, and an OpenSSF Scorecard.
• Open governance: contribution guidelines, code of conduct, security policy, and a public roadmap.
• Telemetry is optional and can be disabled via the linked documentation.

Where to learn more

• AI Engine Docs – detailed API and prompt‑library reference.
• MCP Setup – how to run the Managed Cloud Platform UI.
• CLI docs – command‑line usage examples.
• Support – GitHub Issues, Discussions, and a dedicated Support Guide.

---

The DevOps AI Toolkit brings LLM‑driven intelligence to everyday cloud‑native operations, turning natural‑language requests into concrete Kubernetes actions.