# Software Test Review

Evaluate the quality of TDD tests against slice acceptance criteria, codebase...

- Repository: https://github.com/stencila/stencila (914 stars)
- Category: Testing
- License: Apache-2.0
- Last updated: 2026-11-01
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add stencila/stencila --skill software-test-review -g -a claude-code -y`
Codex: `npx skills add stencila/stencila --skill software-test-review -g -a codex -y`
Cursor: `npx skills add stencila/stencila --skill software-test-review -g -a cursor -y`
Gemini CLI: `npx skills add stencila/stencila --skill software-test-review -g -a gemini-cli -y`
Claude Code plugin marketplace: `/plugin marketplace add stencila/stencila`

## Description

Evaluate the quality of TDD tests against slice acceptance criteria, codebase conventions, and Red-phase execution results, producing a structured review with Accept or Revise recommendations. Use when tests written during the Red phase of red-green-refactor need quality review, checking coverage of acceptance criteria, conformance with codebase test conventions, test quality (naming, assertions, isolation, readability, triviality), edge-case and error-path coverage, and whether Red-phase failures indicate correctly missing implementation. Flags trivial low-value tests that add more maintenance cost than testing value. Discovers codebase conventions independently and produces an actionable review report.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

## Overview

Review tests written during the Red phase of a TDD cycle. This skill evaluates tests across multiple quality dimensions: acceptance-criteria coverage, codebase convention conformance, test quality (including detection of trivial low-value tests), edge-case handling, and whether Red-phase failures indicate correctly missing implementation. The output is a structured review report with an Accept or Revise recommendation.

This skill does not write or modify any code or test files. It only reads, evaluates, and reports.

The review answers one central question: **Are these tests good enough to drive the Green phase of implementation?** Tests that are accepted should be a reliable specification of the slice's expected behavior. Tests that need revision should come back with specific, actionable feedback so the test-creation agent can fix them efficiently.

A corollary question is equally important: **Does every test earn its keep?** Each test carries ongoing maintenance cost, it must be loaded, run, kept passing through refactors, and understood by future developers. A test is only worth that cost if it provides meaningful confidence that a real behavior works correctly. Trivial tests that merely restate the code, assert tautologies, or verify something the type system already guarantees add net-negative value: they cost time to maintain and provide no real safety net. This review must actively identify and flag such tests.

The reviewer must distinguish between acceptance criteria that should drive automated tests and criteria that are better satisfied by implementation or human inspection. Documentation requirements, doc comments, README edits, changelog entries, and similar non-executable deliverables are usually **not** reasons to demand source-inspecting unit tests. Their absence from the test suite is not automatically a coverage gap.

## Required Inputs

This skill requires the following information to operate:

The rest: https://github.com/stencila/stencila/blob/HEAD/.stencila/skills/software-test-review/SKILL.md