Kibi MCP Server

Connect Kibi to Claude, Cursor or any other MCP client and it stops being a tab you switch to. Kibi Cursor plugin - thin adapter for Cursor MCP integration and editor hooks. The kibi mcp server is what makes that connection.

What the server does

Kibi is a repo-local, per-git-branch, queryable knowledge base for software projects. It stores requirements, scenarios, tests, architecture decisions, and more as linked entities, ensuring end-to-end traceability between code and documentation.

• Tracks context across branches — — Every git branch gets its own KB snapshot, preserving context as you switch between features
• Enforces traceability — — Links code symbols to requirements, preventing orphan features and technical debt
• Validates automatically — — Rules catch missing requirements, dangling references, and consistency issues
• Agent-friendly — — LLM assistants can query and update knowledge base via MCP without risking file corruption
• Guides semantic modeling — — The MCP server can inspect prose requirements and suggest strict facts or reusable predicate facts before agents treat text as machine-checkable knowledge

Installation

args on npm is all you need. Most clients run it directly, so configuration is a few lines and a restart.

Available tools

The toolset is worth reading before you wire it up, because it tells you what the integration is really for:

• req — Software requirements specifying functionality or constraints
• scenario — BDD scenarios describing user behavior (Given/When/Then)
• test — Executable unit, integration, or e2e test cases
• fact — Atomic domain facts and invariants. Supports a strict lane for contradiction-sensitive modeling and a context lane (observation, meta) for
• adr — Architecture Decision Records documenting technical choices
• flag — Runtime or config gates (feature flags, kill-switches)
• event — Domain or system events published/consumed by components
• symbol — Abstract code symbols (functions, classes, modules)
• Agent-friendly — LLM assistants can query and update knowledge base via MCP without risking file corruption
• Event — Driven Architecture** — Map who publishes and consumes each domain event. Trace event flows and identify couplings across the system
• Branch — Local Memory** — Every git branch keeps its own KB snapshot. Switch contexts without losing traceability or polluting other branches
• kibi-core — Prolog-based knowledge graph that tracks entities across branches

Credentials and setup notes

• SWI-Prolog 9.0+ — Kibi's knowledge graph runs on Prolog

Worth knowing first

• It runs with your machine's permissions. That is convenient and also the reason to think about what you point it at before you approve a tool call.
• With 14 tools registered it takes up a noticeable share of the context window; turn it off in projects that never touch Kibi.
• Keep per-call confirmation enabled while you learn its behaviour; it is the cheapest safeguard you have.

Where it fits

Among the developer tooling options, the useful question is rarely "what can it do" but "what does it cost you to run" — permissions, credentials, and how much of your context its toolset consumes. Kibi's toolset — req, scenario, test and 11 more — is a fair guide to whether it matches your workflow. It is maintained by looted; worth a glance at recent repository activity before you build anything load-bearing on it.

SyncDev reviews every entry in this directory against the project's own documentation before publishing, and revisits them as servers change.