4meta5

suggest-tests

0
0
# Install this skill:
npx skills add 4meta5/skills --skill "suggest-tests"

Install specific skill from multi-skill repository

# Description

Recommend unit tests to add based on git diff changes. Analyzes modified code and prioritizes by risk level.

# SKILL.md


name: suggest-tests
description: Recommend unit tests to add based on git diff changes. Analyzes modified code and prioritizes by risk level.
category: testing
disable-model-invocation: false
user-invocable: true
allowed-tools: Read, Grep, Glob, Bash


Suggest Tests from Git Diff

Inspect the repository for code changes and recommend test additions or updates.

Inputs

  • diffScope: 'branch' or 'commit'
  • referenceBranchOrCommit (optional): The reference branch or commit to compare from

Procedure

Phase 1 - Parse the Diff

  1. Get all changes in this branch compared to the reference (default branch or specified commit)
  2. Identify changed or added functions, methods, classes, constants, or modules
  3. Ignore comment-only or formatting changes

Phase 2 - Analyze the Changes

  1. For each affected symbol:
  2. Determine change type: added | modified | deleted | refactored
  3. Detect whether the logic, signature, or control flow changed
  4. Cross-reference coverage data if provided
  5. Prioritize recommendations:
  6. HIGH: New code or core logic changes with missing or low test coverage
  7. MEDIUM: Modified parameters, return types, or conditionals
  8. LOW: Trivial changes with existing adequate coverage
  9. Flag risky changes (input validation, deserialization, external calls, etc.)

Phase 3 - Output Recommendations

Return Markdown structured as follows:

# Test Recommendations

## Risk Level High
**functionName**
- location: src/path/to/file.ts
- change type: modified
- reason: Description of why this needs tests
- suggested tests:
  - Test case description 1
  - Test case description 2

## Risk Level Medium
...

## Risk Level Low
...

Constraints

  • Analyze only within the given code. Do not invent missing context or external APIs
  • Be deterministic and concise
  • Return only recommendations

Skill Chaining

After Analysis

Chain To When Action
unit-test-workflow HIGH risk identified Generate comprehensive tests
property-based-testing Serialization patterns Add PBT coverage

Chains From

Source Condition
tdd After REFACTOR phase
workflow-orchestrator Code review context

Terminal Chain

After completion: repo-hygiene (clean analysis artifacts)

Testing Pipeline Position

suggest-tests is step 2 in the testing pipeline:

tdd → suggest-tests → unit-test-workflow → property-based-testing → repo-hygiene
         ↑
     (you are here)

# Supported AI Coding Agents

This skill is compatible with the SKILL.md standard and works with all major AI coding agents:

Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.