register opus 5.5 and use it for automations

[damienrj]

Summary

Category: improvement

User Impact: Claude Opus 5.5 shows up at the top of the Opus group in the model picker, and the automation builder runs on Opus 5.5 instead of Opus 4.6.

Problem: Goose model IDs such as goose-claude-opus-5-5 were already named and recommended correctly, but models within one family were sorted alphabetically by label, so older Opus versions appeared above Opus 5.5. The automation builder was also pinned to goose-claude-4-6-opus.

Solution: Sort models within the same family newest version first, as GPT models already were. Move the automation builder to databricks-claude-opus-5-5 in both the renderer and the Rust request allowlist; the old ID is now rejected. kgoose passes a Databricks preferredModel.name through unchanged, so it must be an existing Databricks serving endpoint; databricks-claude-opus-5-5 is the Opus 5.5 endpoint.

File changes

src/features/providers/lib/modelRecommendations.ts
Orders models within one family newest version first.

src/features/providers/lib/modelRecommendations.test.ts
Covers Opus 5.5 naming, ordering above Opus 4.8 and 4.6, recommendation, and featuring.

src/features/providers/lib/humanizeModelId.test.ts
Covers raw Opus 5.5 IDs.

src/features/automations/api/automationBuilder.ts
Requests databricks-claude-opus-5-5 for the automation builder.

src/features/automations/api/automationBuilder.test.ts
Updates the expected builder model.

src-tauri/src/commands/automations/sanitize.rs
Allows only databricks-claude-opus-5-5 as the builder's preferred model and tests that goose-claude-4-6-opus is now rejected.

Related issue

N/A

Testing

• just check passes; the provider and automation Vitest suites pass.
• The automation sanitize Rust tests pass locally (16 tests) with --features block-automations. CI does not currently run this module's tests because it does not enable that feature.
• End-to-end automation creation against kgoose has not yet been exercised with the new model name.

[morgmart]

🤖 Automated code review

COMMENT: no publishable inline defects were found. Morgan should decide whether backend support for the automation builder's new required model must be confirmed before merge. Several GitHub checks were still in progress when captured and continue to govern merge readiness.

Deterministic publication result: 0 blocking and 0 non-blocking inline finding(s) publishable; 0 duplicate(s) suppressed; 0 blocking screenshot-evidence requirement(s) in this review body.

[morgmart]

🤖 Automated code review

APPROVE: the full exact PR comparison is coherent and has sufficient discriminating coverage for the automation model identifier and model recommendation ordering changes. No publishable findings, duplicate suppressions, or escalations remain. Supplied GitHub check evidence is structurally valid; all captured check runs completed successfully, while required checks still independently govern merge readiness.

Deterministic publication result: 0 blocking and 0 non-blocking inline finding(s) publishable; 0 duplicate(s) suppressed; 0 blocking screenshot-evidence requirement(s) in this review body.

Pending checks: 1 check(s) are not complete.

This approval reflects the completed code review only; merge readiness remains governed by the repository's required checks.