Archgate

Archgate is a command-line tool designed to provide enterprise-grade linting and guardrails for teams working with AI-driven development. It addresses the challenge of maintaining architectural consistency and code quality by leveraging Architecture Decision Records (ADRs) that are both human-readable and machine-enforceable. ADRs are documented in Markdown files with YAML frontmatter, describing architectural decisions in plain language, and are paired with companion TypeScript files that define automated compliance checks.

The platform enables users to write rules in TypeScript, which Archgate then executes against the codebase to detect and report violations, including the specific file paths and line numbers where issues occur. This allows teams to ensure that coding conventions and architectural guidelines are consistently followed. When running the 'archgate check' command, the tool loads all ADRs marked with rules and executes their corresponding TypeScript rules files, reporting any violations and returning an exit code to indicate compliance status. An exit code of 0 signifies compliance, while 1 indicates violations have been found.

Archgate integrates with continuous integration (CI) systems, including GitHub Actions and GitLab CI, by using exit codes to block merges when rules are violated. This ensures that non-compliant code does not enter the main codebase. The tool also offers editor plugins, currently in beta, that extend its functionality to AI agents and development environments. These plugins allow AI agents to read ADRs before generating code, validate code after writing, and capture new patterns automatically. Specific integrations mentioned include plugins for Claude Code, Cursor IDE, VS Code, GitHub Copilot CLI, and a workflow that provides AI agents with structured skills such as developer, architect, quality-manager, adr-author, and onboard.

Archgate is intended for engineering teams seeking to automate and enforce architectural decisions and code quality standards, particularly in environments where AI agents contribute to codebases. The tool is delivered as a CLI application, with additional functionality available through editor plugins and integration guides for various development tools.