feat(provider): add alibaba_coding provider with Anthropic-compatible model catalog

[amitksingh1490]

Summary

Add a new alibaba_coding LLM provider entry to the provider catalog so Forge can route Anthropic-compatible requests through Alibaba Coding endpoints and expose its model lineup.

Context

Forge’s provider catalog did not include Alibaba Coding, which prevented users with ALIBABA_CODING_API_KEY and compatible endpoint configuration from selecting Alibaba Coding models.

Changes

• Added a new provider definition with id: alibaba_coding
• Configured Anthropic-compatible response mode and /messages endpoint templating
• Added provider auth and URL variable wiring (ALIBABA_CODING_API_KEY, ANTHROPIC_COMPACT_URL)
• Registered 8 models with metadata and capability flags:
  • qwen3.5-plus
  • kimi-k2.5
  • glm-5
  • MiniMax-M2.5
  • qwen3-max-2026-01-23
  • qwen3-coder-next
  • qwen3-coder-plus
  • glm-4.7

Key Implementation Details

The new provider follows the existing provider schema pattern in provider.json, including tools_supported, supports_parallel_tool_calls, supports_reasoning, and input_modalities flags for each model.

Use Cases

• Use Alibaba Coding hosted Qwen/GLM/Kimi models through existing provider selection flows
• Enable tool-calling and reasoning-capable model choices under a single provider config
• Keep configuration compatible with Anthropic-style message APIs already used in Forge

Testing

• Validated JSON schema/syntax locally:
  • jq empty crates/forge_repo/src/provider/provider.json

Links

• Related issues: N/A