Harden Sage editing and deterministic QA workflows

[Valorith]

What changed

• added deterministic Sage NPC appearance, animation, texture, spawn, door, and model-viewer validation workflows
• added the standalone Model Review workspace with centered whole-model and face-focus framing
• integrated Model Review with the Sage QA runner, repeatable screenshots, visual invariants, and HTML/JSON reports
• hardened spawn and door lifecycle behavior and added focused Playwright coverage
• added model-scoped semantic Iksar head-orientation validation that fails closed without applying a destructive rig transform
• fixed Evolving Items navigation to stay inside the application

Why

Sage visual and editor regressions were difficult to reproduce reliably because validation depended on manual inspection, fixed timing, and incomplete model/appearance coverage. The new tooling gives developers a repeatable production-surface workflow for verifying appearance variants, animations, model framing, spawn edits, door placement, and resource stability.

Impact

Developers can now inspect models as they render in Spire, focus on faces, exercise editor lifecycles, and capture stable evidence through one QA runner. Runtime changes are scoped and validated to avoid changing correctly authored models.

Root cause

The affected workflows lacked stable readiness contracts and semantic visual invariants. In particular, prior head-orientation checks covered texture state and gross bounds but could not distinguish a valid Iksar head from a damaging whole-head rotation.

Validation

• 73/73 Sage QA unit tests
• full frontend production build, including the 2,942-module EQ Sage embed and Vue distribution
• focused production Model Review profile: Iksar base front, base face, and alternate face all passed twice
• semantic Iksar cranium-to-jaw separation checks passed at 0.44 and 0.33 versus a 0.04 minimum
• cached diff and whitespace checks passed

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 2a3a120b-ff99-44d5-9539-06111414dc17

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch codex/sage-reliability

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}