Request: add explicit pre-flight warnings + agent best-practices for complex visual/3D workflows #43

Summary

When using agentic coding workflows for complex visual projects (Three.js/Vite, scene-heavy UIs), users need an explicit pre-flight warning and setup checklist before agents begin making scene changes.

Problem

In ambiguous scene contexts, agents may make valid-but-wrong assumptions (placement, camera/framing, flow mapping, asset control hierarchy). This can cause large time loss and repeated rework, especially when no project-specific guardrails are initialized first.

Requested product behavior

Please add a default precautionary warning / best-practices notice before major scene edits, including:

1. Confirm project intent before coordinate/placement edits.
2. Recommend enabling a dev scene inspector first (object selector, transform/material/animation, keyframe, projection, source-apply).
3. Warn that ambiguous instructions (e.g., "build a 1000m curve camera") require explicit constraints.
4. Encourage diagnostics-first steps (visibility/frustum/material/clip state) before mutation.
5. Suggest setup of explicit logging/functions for fallback behavior.
6. Warn that runtime edits should translate back to source (JS/CSS/HTML), not remain ephemeral.
7. Include a legal/operational disclaimer for potential productivity loss from ambiguous agent assumptions.

Why this matters

Professional teams often avoid these pitfalls by preconfiguring IDE/engine/pipeline tools. Agentic workflows are different and need first-class safety rails to avoid expensive regressions in real projects.

Suggested UX

A lightweight "Complex Visual Workflow Detected" prompt with checkboxes:

• Require coordinate-confirmation for spawn/root transform changes
• Enable diagnostics mode first
• Initialize scene inspector utilities
• Enable guarded render-loop integration for tool failures
• Enable one-click source sync for interactive edits

This would materially improve reliability and user trust for complex scene development.