AgentSkillsCN

ask-laravel-mechanic

在零数据丢失政策下维护Laravel应用，同时进行MongoDB与SQL数据库的调试工作。

SKILL.md
--- frontmatter
name: ask-laravel-mechanic
description: Laravel maintenance with Zero Data Loss policy, Mongo/SQL debugging.
triggers: ["fix n+1 query", "debug laravel queue", "check log files", "restore soft deleted"]

<critical_constraints> ❌ NO migrate:fresh on prod/staging → destroys DB ❌ NO db:seed on prod → overwrites data ❌ NO Model::truncate() on prod without backup ❌ NO migrate:reset on prod ✅ MUST run php artisan env before dangerous commands ✅ MUST use migrate --pretend when unsure ✅ MUST restart queue after code deployment </critical_constraints>

<safety_commands> ✅ Allowed: migrate (forward only) ✅ Safe test: migrate --pretend (shows SQL without running) ✅ Cache clear: optimize:clear ✅ Queue restart: queue:restart (after deploy!) </safety_commands>

<soft_delete_restore>

php
User::withTrashed()->find($id)->restore();

</soft_delete_restore>

<debugging> ## Database Inspection - SQL: `php artisan model:show User` - Mongo: `php artisan tinker --execute="dump(App\Models\User::first()->getAttributes())"`

N+1 Prevention

php
// In AppServiceProvider::boot()
Model::preventLazyLoading(!app()->isProduction());

Common Errors

  • •"member function on null" → Mongo relation with SQL syntax, check with()
  • •"MongoDB... not found" → wrong namespace, check composer.json </debugging>

<log_analysis>

bash
# Single channel
tail -n 50 storage/logs/laravel.log

# Daily channel
tail -n 50 storage/logs/laravel-$(date +%Y-%m-%d).log

# Search with context
grep -C 5 "User ID 505" storage/logs/laravel.log

Keywords: local.ERROR, QueryException, ModelNotFound, MassAssignmentException </log_analysis>

<queue_forensics>

  • •Status: queue:monitor default
  • •Failed: queue:failed
  • •Retry one: queue:retry <UUID>
  • •Flush all: queue:flush (careful!) </queue_forensics>