post-change-validation

name	post-change-validation
description	Automatic validation workflow after code changes (loads appropriate skill)
license	MIT
compatibility	opencode
metadata	{"audience":"developers","workflow":"automation"}

What I do

After you've made code changes, I guide you through automatic validation:

• Detect change scope - Determine what was modified
• Recommend validation - Suggest which skill to load based on changes
• Execute validation - Run the appropriate make target
• Report results - Show pass/fail status and next steps

When to use me

Use this skill after any code changes:

1. After writing new functions/classes
2. After modifying existing code
3. After refactoring
4. After fixing bugs
5. Before committing changes

How I work

I analyze your changes and recommend:

For code changes only

→ Load lint-project skill

• Run: make check
• Validates: Formatting, types, linting

For code + test changes

→ Load test-project skill then lint-project

• Run: make test then make check
• Validates: Tests + coverage + linting

For major changes or PR preparation

→ Load check-project skill

• Run: make check then make test
• Full comprehensive validation

For dependency changes

→ Load lint-project skill

• Run: make check
• Validates: Lock file + dependency audit

Example workflow

You: [Make code changes]
Agent: [Detect changes and load this skill]
Validation Skill: Analyzing changes...
→ Code modified: app.py
→ Tests modified: test_app.py
Recommendation: Load test-project skill
Agent: [Load test-project]
Test Skill: Running make test...
Results: ✅ All tests pass (89% coverage)
Agent: Ask what to do next

What triggers recommendations

Just code changes → lint-project

Modified: app.py, utils.py
→ Recommend: lint-project

Code + Tests → test-project

Modified: app.py, test_app.py
→ Recommend: test-project

Dependencies changed → lint-project

Modified: pyproject.toml
→ Recommend: lint-project

Complex changes → check-project

Modified: 5+ files, including tests
→ Recommend: check-project

Integration with agent protocol

This skill works with your agent-interaction rules:

1. Agent makes code changes
2. Agent detects this skill is available
3. Agent recommends loading this skill to user
4. User chooses validation approach
5. Agent loads recommended skill (or alternative)
6. Agent runs make target via loaded skill
7. Agent asks user what to do next

Quick references

Shallow validation (lint only):

• Use: lint-project
• Time: ~1 minute
• Good for: Type/format fixes

Deep validation (lint + test):

• Use: test-project
• Time: ~2-3 minutes
• Good for: New features, tests

Complete validation (full pipeline):

• Use: check-project
• Time: ~3-4 minutes
• Good for: PRs, releases

Pro tips

• Load this skill after each meaningful code change
• It will guide you to the right validation tool
• Trust the recommendations but override if needed
• Use check-project when in doubt

---

Related Skills: lint-project, test-project, check-project