feat: add alibaba coding plan as provider

[joshuachong]

Summary

• Reraising PR from feat: add alibaba coding plan as provider #849
• Fixes Integrate Alibaba Coding Plan Models #418 and replicates feat: add Alibaba Coding (DashScope Intl) provider with 8 models #747 for correct integration, documentation and model specs.
• Adds Alibaba Coding Plan (International) subscription as a model provider into OpenFang provider list. Coding Plan is a subscription based and not pay-per-token service
• Integrating using reference from Alibaba Coding Plan's OpenClaw integration

Notes:

• API Key format: sk-sp-xxxxx (plan-specific, different from pay-as-you-go sk-xxxxx)
• Base URL: https://coding-intl.dashscope.aliyuncs.com/v1
• Subscription: $50/month Pro plan with quota limits (90K requests/month, 45K/week, 6K per 5 hours)

Note: Additional linting errors are coming from drivers/copilot.rs changes. Not changing untouched files for now.

Model Specifications

Model ID	Display Name	Tier	Context	Max Output	Vision	Tools	Cost
alibaba-coding-plan/qwen3.6-plus	Qwen 3.6 Plus	Smart	1,000,000	65,536	✅	✅	$0
alibaba-coding-plan/qwen3.5-plus	Qwen 3.5 Plus	Smart	1,000,000	65,536	✅	✅	$0
alibaba-coding-plan/qwen3-max-2026-01-23	Qwen 3 Max	Frontier	262,144	65,536	❌	✅	$0
alibaba-coding-plan/qwen3-coder-plus	Qwen 3 Coder Plus	Smart	1,000,000	65,536	❌	✅	$0
alibaba-coding-plan/qwen3-coder-next	Qwen 3 Coder Next	Frontier	262,144	65,536	❌	✅	$0
alibaba-coding-plan/glm-5	GLM-5 (Zhipu)	Frontier	202,752	32,768	❌	✅	$0
alibaba-coding-plan/glm-4.7	GLM-4.7 (Zhipu)	Smart	202,752	32,768	❌	✅	$0
alibaba-coding-plan/kimi-k2.5	Kimi K2.5 (Moonshot)	Smart	262,144	32,768	✅	✅	$0
alibaba-coding-plan/MiniMax-M2.5	MiniMax M2.5	Smart	204,800	32,768	❌	✅	$0

Changes

• Added Alibaba Coding Plan as constants and model_catalog
• 9 model entries added as part of Alibaba Coding Plan support
• Added pattern matching for Alibaba Coding Plan on metering since this is a subscription-based model and not pay-per-token
• Updated documentation for new provider

Testing

• cargo clippy --workspace --all-targets -- -D warnings passes
• cargo test --workspace passes
• Live integration tested (if applicable)

Security

• No new unsafe code
• No secrets or API keys in diff
• User input validated at boundaries

[joshuachong]

Note linting errors from crates/openfang-runtime/src/drivers/copilot.rs that are not part of this change:

warning: constant `DEVICE_FLOW_POLL_INTERVAL` is never used
  --> crates/openfang-runtime/src/drivers/copilot.rs:58:7
   |
58 | const DEVICE_FLOW_POLL_INTERVAL: Duration = Duration::from_secs(5);
   |       ^^^^^^^^^^^^^^^^^^^^^^^^^
   |
   = note: `#[warn(dead_code)]` (part of `#[warn(unused)]`) on by default

warning: field `fetched_at` is never read
   --> crates/openfang-runtime/src/drivers/copilot.rs:141:5
    |
139 | struct CachedModels {
    |        ------------ field in this struct
140 |     models: Vec<String>,
141 |     fetched_at: Instant,
    |     ^^^^^^^^^^
    |
    = note: `CachedModels` has a derived impl for the trait `Clone`, but this is intentionally ignored during dead code analysis

warning: field `refresh_token_expires_in` is never read
   --> crates/openfang-runtime/src/drivers/copilot.rs:172:5
    |
164 | struct OAuthTokenResponse {
    |        ------------------ field in this struct
...
172 |     refresh_token_expires_in: Option<i64>,
    |     ^^^^^^^^^^^^^^^^^^^^^^^^

warning: function `prompt_line` is never used
   --> crates/openfang-runtime/src/drivers/copilot.rs:785:4
    |
785 | fn prompt_line(prompt: &str) -> Result<String, String> {
    |    ^^^^^^^^^^^

[jaberjaber23]

Thanks @joshuachong — Alibaba Dashscope as a provider is a welcome addition. Verified the base URL (https://coding-intl.dashscope.aliyuncs.com/v1) is the official Dashscope-international endpoint, Bearer auth via ALIBABA_CODING_PLAN_API_KEY matches their docs.

Before merge:

1. Clippy/lint fallout — your PR notes mention lint warnings in copilot.rs that are not from your changes. Rebase on current main (post-fix(deps): upgrade wasmtime 41->43 and rumqttc 0.24->0.25 to resolve active CVEs #1041 which also touched drivers) and run cargo clippy --workspace -- -D warnings locally. If the noise persists, flag it and I'll look.
2. Parallel-test flakiness — your driver tests use std::env::set_var / remove_var which race across cargo test threads. Either wrap with a serial_test attribute or use a Mutex<()> guard for the API-key env tests.
3. Doc/code model count sync — docs/providers.md lists ~62 models for Alibaba; please confirm builtin_models() returns exactly that number so the numbers don't drift.

Once those are in and CI is green, this is ready to merge.