feat(llm): add multi-LLM provider support with API key validation (#4)

* feat(llm): add multi-LLM provider support with API key validation

## Features
- Add 6 LLM providers: DeepSeek, MiniMax, ZhiPu, Qwen, OpenAI, Ollama
- Add --llm CLI option for non-interactive provider selection
- Add Docker Volume mounting strategy for categorization.py patch
- Add optional API key validation before saving

## Code Review Fixes
- H1: Add providers.test.ts with 22 test cases
- H2: Add --llm option tests in install.test.ts
- H3: Auto-update .gitignore and show security warning for .env
- H4: Fix Ollama envKey empty string bug in generateProviderEnv
- M1: Handle invalid provider in non-interactive mode
- M2: Add Ollama support in categorization.py
- M3: Use getMcpEnvForProvider in showMcpConfig
- L1: Auto-generate PROVIDER_CHOICES from LLM_PROVIDERS
- L2: Add validateApiKey function with tests

## Files
- cli/src/lib/providers.ts (new)
- cli/templates/patches/categorization.py (new)
- cli/templates/docker-compose.full.yml (new)
- cli/tests/providers.test.ts (new)
- cli/src/commands/install.ts (modified)
- cli/src/index.ts (modified)
- cli/tests/install.test.ts (modified)
- .env.example (modified)

Tests: 119 passed

* chore: add self-hosted omp templates (dogfooding)

Add generated templates for self-hosting openmemory-plus:
- .augment/commands/memory.md - Augment IDE memory command
- .augment/skills/memory-extraction/ - Memory extraction skill
- _omp/commands/ - Shared memory command
- _omp/skills/ - Shared skills
- _omp/workflows/memory/ - Memory workflow (9 steps)

* feat: merge main with LLM provider support

- Merge main branch MCP auto-config features
- Re-add multi-LLM provider selection (DeepSeek, MiniMax, ZhiPu, Qwen, OpenAI, Ollama)
- Add patches directory copy for categorization.py
- Add API key validation and .env generation
- All 128 tests passing

Author: Alenryuichi

.augment/commands/memory.md

@@ -0,0 +1,6 @@
+---
+name: 'memory'
+description: '记忆管理 - 统一管理项目级和用户级记忆'
+---
+
+IT IS CRITICAL THAT YOU FOLLOW THIS COMMAND: LOAD the FULL @{project-root}/_omp/workflows/memory/workflow.md, READ its entire contents and follow its directions exactly!