docs: Unify guidance — AGENTS.md as source of truth

Establish clear governance hierarchy:
1. AGENTS.md — AI coding constitution, authoritative
2. CLAUDE.md — Project context, supplements AGENTS.md
3. Code — Implementation, must comply with AGENTS.md

Changes:
- AGENTS.md: Add explicit declaration that this is the authoritative source
- CLAUDE.md: Add governance statement, redirect database config to AGENTS.md
- Quick Start: Remove stale/incorrect database credentials, point to appsettings.Development.json

Prevent guidance fragmentation:
- Do not add conflicting rules to CLAUDE.md
- All harnesses and procedures go to AGENTS.md
- Supplementary info only in CLAUDE.md (status, architecture, overview)

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>

AGENTS.md

@@ -1,4 +1,13 @@
-# K-ArtSell Aegis AI Coding Constitution v12.0
+# K-ArtSell Aegis AI Coding Constitution v16.0
+
+**AGENTS.md IS THE AUTHORITATIVE SOURCE FOR ALL ENGINEERING DECISIONS.**
+
+This document supersedes conflicting guidance in CLAUDE.md, MEMORY, or prior conversations. When in doubt, follow AGENTS.md.
+
+**Hierarchy:**
+1. **AGENTS.md** — This file (AI coding constitution, rules, harnesses)
+2. **CLAUDE.md** — Supplements AGENTS.md (project context, architecture, status)
+3. **Code** — Implementation (must comply with AGENTS.md)
 
 ## Default execution procedure