Add AIgateway as LLM provider

[Rakesh1002]

Description

Adds AIgateway as a first-class LLM provider, following the existing Tensorix pattern (an OpenAI-compatible aggregator):

• core/llm/llms/AIgateway.ts — provider class extending OpenAI with apiBase: https://api.aigateway.sh/v1
• Registered in core/llm/llms/index.ts, core/llm/autodetect.ts (templating / images / parallel lists)
• packages/openai-adapters — OpenAI-compatible adapter mapping + schema literal
• gui/src/pages/AddNewModel/configs/providers.ts — Add New Model entry
• extensions/vscode/config_schema.json — provider enum, description, apiKey-required condition, embeddings enums
• docs/customize/model-providers/more/aigateway.mdx + sidebar entry in docs/docs.json

AIgateway aggregates 1,000+ models from 85+ labs (Claude, GPT, Gemini, Kimi, GLM, Llama, ...) behind one OpenAI-compatible endpoint, with model IDs as provider/slug pairs. Default model is zai-org/glm-5.3-flash (1.3M context, tool use, vision).

Happy to adjust anything — naming conventions, default model, docs wording.

[github-actions]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[Rakesh1002]

The jetbrains-tests job is failing on its environment setup — Downloading FFmpeg 7.1 for Linux X64 → xz: (stdin): File format not recognized → tar: exiting now — before any test or build step runs. Looks like a broken toolchain download rather than anything in this diff (all other 47 checks pass, and the diff doesn't touch JetBrains sources). Could someone with rerun rights re-trigger it?