TDDGuard
title: TDD Guard type: tool created: 2026-05-16 last_updated: 2026-05-16 related: ["Playradar", "radar/techniques/LLMTDDLoop"] sources: ["https://github.com/nizos/tdd-guard"] radar_quadrant: Tools radar_ring: Assess radar_position: inner
A Claude Code plugin that enforces Test-Driven Development discipline by blocking implementation unless a failing test exists and preventing the AI from writing code that exceeds current test requirements.
Claude Code, left unconstrained, tends to implement beyond the current test scope: writing helper functions not yet tested, adding error handling not covered by a failing test, or skipping the failing-test step entirely. TDD Guard intercepts these patterns at the Claude Code hook level and blocks non-compliant actions before they execute.
| Rule | Description |
|---|---|
| Test-first | Blocks implementation without a prior failing test |
| Minimal implementation | Prevents code that goes beyond current test scope |
| Lint integration | Enforces refactoring via existing linting rules during the refactor phase |
| Customisable rules | Validation rules are configurable per project |
The plugin operates mid-session and can be activated or deactivated without restarting Claude Code.
Vitest, Jest, Storybook, pytest, PHPUnit, Go (testing package), Rust (cargo test), RSpec, Minitest.
/plugin marketplace add nizos/tdd-guard
/plugin install tdd-guard@tdd-guard
/tdd-guard:setup
Requires Node.js 22+.
2,100+ GitHub stars, 77 releases as of May 2026 (latest: v1.6.8). Written primarily in TypeScript with test framework adapters in Ruby, Go, Rust, PHP, and Java. MIT licensed.
TDD Guard sits in the Assess ring of the Tools quadrant, at inner position. First studied via the GitHub repository on 2026-05-16; no personal use. The tool is active, well-maintained, and directly applicable to Claude Code-driven development workflows. Inner position reflects that installation friction is minimal (three commands), it supports the test frameworks in active use (pytest, Go), and the problem it solves — AI code drift from TDD discipline — is a real and observable behaviour. The remaining gate before Trial is a first use on an active project with Claude Code.
