Add Requesty as an LLM provider

[Thibaultjaigu]

Adds Requesty as an LLM/embedding provider alongside the existing OpenRouter and Ollama integrations.

Requesty (https://router.requesty.ai/v1) is an OpenAI-compatible router, so the integration mirrors the existing OpenRouter provider:

• New RequestyClient / RequestyClientManager in server/integrations/requesty.py (async httpx client, Bearer REQUESTY_API_KEY auth, same provider/model naming e.g. openai/gpt-4.1-mini, same JSON-extraction helpers). verify_api_key uses the OpenAI-compatible /models endpoint.
• Registered in integrations/__init__.py exports, config/settings.py (LLM_PROVIDER=requesty, REQUESTY_BASE_URL, default https://router.requesty.ai/v1), and the http_server client-manager factory + /api/auth/register validation branch.
• Added a synchronous skill util (SKILL/simplemem-skill/src/utils/requesty.py) and references/requesty-guide.md mirroring the OpenRouter skill files.
• Documented in .env.example, README, and the skill READMEs.
• Changes are applied to both the top-level MCP/ + SKILL/ trees and their simplemem/integrations/ mirrors to keep them in sync.

Set LLM_PROVIDER=requesty and register with a Requesty API key (https://app.requesty.ai/api-keys) to use it. Docs: https://docs.requesty.ai

Verification:

• py_compile passes on all new/edited Python files.
• Live test against https://router.requesty.ai/v1 through the new RequestyClient: verify_api_key() returned True, and chat_completion(openai/gpt-4o-mini) returned "ok".

I work at Requesty. This mirrors the existing OpenRouter provider as closely as possible. Happy to adjust or close it if it's not a fit.