mcpField
T
test-tier-selection

test-tier-selection

github.com/laurigatesTesting & quality5958 installsUpdated 1d ago
Repository

About test-tier-selection

test-tier-selection is a guidance skill for developers choosing test scope based on the type and size of a change. Its decision matrix and escalation signals indicate when to run unit, integration, or end-to-end tests.

Key features

  1. 01Select test tiers by change type
  2. 02Provide criteria for escalating test scope
  3. 03List commands for each tier
  4. 04Identify when to consult related agents

Use cases

  1. 01Choose tests for a small code change
  2. 02Set test scope for feature or API changes
  3. 03Plan testing before a commit or PR

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/laurigates/claude-plugins --skill test-tier-selection

After installing, type/test-tier-selectionin Claude Code to confirm it loaded.

Details

From the project README

Test Tier Selection

Automatic guidance for selecting appropriate test tiers based on change context and scope.

Test Tier Definitions

TierDurationScopeWhen to Run
Unit< 30sSingle function/moduleAfter every code change
Integration< 5minComponent interactionsAfter feature completion
E2E< 30minFull user flowsBefore commit/PR

Decision Matrix

Change Type → Test Tier
Change TypeUnitIntegrationE2E
Single function fixRequiredSkipSkip
New feature (1 file)RequiredRequiredSkip
Multi-file featureRequiredRequiredRequired
RefactoringRequiredRequiredOptional
API changesRequiredRequiredRequired
UI changesRequiredOptionalRequired
Bug fix (isolated)RequiredOptionalSkip
Database changesRequiredRequiredRequired
Config changesRequiredRequiredOptional

Escalation Signals

Escalate to Integration when:

  • Changes span multiple files
  • Business logic affected
  • Service boundaries modified
  • Database queries changed

Escalate to E2E when:

  • User-facing features modified
  • Authentication/authorization changes
  • Critical path functionality
  • Before creating PR

Commands by Tier

# Tier 1: Unit (fast feedback)
/test:quick

# Tier 2: Integration (feature completion)
/test:full --coverage

# Tier 3: E2E (pre-commit)
/test:full

Agent Consultation Triggers

Consult test-architecture agent when:

  • New feature module created
  • Coverage drops > 5%

  • 3 flaky tests detected

  • Framework questions arise
  • Test strategy needs adjustment

Consult test-runner agent when:

  • Need test execution with analysis
  • Multiple failures to diagnose
  • Want concise failure summary

Consult system-debugging agent when:

  • Integration test failures with unclear cause
  • Environment/timing issues
  • Flaky tests related to concurrency

Quick Reference

After Small Change
1. Run /test:quick
2. If pass: Continue working
3. If fail: Fix immediately
After Feature Completion
1. Run /test:full --coverage
2. Check coverage targets met
3. If gaps: /test:consult coverage
Before Commit/PR
1. Run /test:full
2. All tiers must pass
3. Review coverage report
For New Features
1. /test:consult new-feature
2. Write tests (TDD)
3. Run /test:quick during development
4. Run /test:full before PR

Activation Triggers

This skill auto-activates when:

  • User mentions "test", "run tests", "testing"
  • After code modification by Claude
  • During TDD workflow
  • When /test:* commands invoked
  • When discussing test strategy

Comments

Markdown supported