tenets, created by James Dunn (jddunn), is an MCP server that maps project rules to AI coding sessions. The server supplies live, project-specific guidance so code generated by an assistant adheres to established standards. Key functions include a searchable rule library, context injection into model sessions, and dynamic rule toggles. It targets software engineers and technical architects who need predictable, project-aligned outputs from AI coding assistants.

Tenets acts as a protocol-native bridge between standards and model sessions

The server is built specifically for the Model Context Protocol and exposes tenets to MCP-compatible clients, which lets an AI client query project rules during an editing session. That protocol focus distinguishes it from ad hoc system prompts by providing a structured endpoint the model can request, rather than requiring rule copy-paste into each prompt.

Rule management and persistence let teams keep a single source of truth

The tool offers full CRUD for coding principles and persists rules in a local configuration, which ensures the rule set remains available across sessions. Administrators can add, update, or remove tenets without restarting a session, and the storage format is a local JSON file, enabling project-level organization using tags or names.

It integrates into developer workflows but requires specific runtime components

Installation requires a Node.js environment (version 18 or higher recommended) and an MCP-compatible client such as Claude Desktop to consume context. Setup options include cloning the repository or invoking the package via npx, so the server fits into existing developer toolchains that support MCP endpoints.

Practical behavior and operational limits you should expect

The server injects tenets into AI sessions, which reduces policy drift for model outputs, yet the AI client typically processes that context remotely, so verification remains necessary for sensitive code. The project is noted among early MCP adopters on GitHub, showing community interest but also implying an active-development status rather than a mature enterprise platform.

Practical governance for teams embedding AI in development workflows

For development teams that require repeatable, project-specific constraints during AI-assisted coding, the server provides a pragmatic governance layer tied to a protocol clients already use. Expect a modest setup effort to run the Node.js server and maintain rule libraries; combine the server with human review or CI checks to catch context-misaligned suggestions the model may still produce.