feat: sticky header with shared model selector

.agents/archive/descope-self-hosting.md

@@ -555,10 +555,10 @@ For each of these files, find and remove or update Ollama/self-hosting/Docker me
 
 - `.agents/context/architecture.md` — remove Ollama from architecture diagram
 - `.agents/context/glossary.md` — remove Ollama model references
-- `.agents/context/decisions/002-vercel-ai-sdk.md` — remove Ollama from supported providers list
-- `.agents/context/decisions/004-exa-vs-tavily.md` — remove "self-hosters" references
-- `.agents/context/research/competitive-feature-analysis.md` — remove self-hosting as competitive advantage
-- `.agents/context/research/tool-calling-infrastructure.md` — remove Ollama mentions
+- `.agents/research/tech-stack-evaluation.md` — remove Ollama from supported providers list
+- `.agents/research/tool-calling-infrastructure.md` — remove "self-hosters" references
+- `.agents/research/competitive-feature-analysis.md` — remove self-hosting as competitive advantage
+- `.agents/research/tool-calling-infrastructure.md` — remove Ollama mentions
 - `.agents/plans/thinking-reasoning-configuration.md` — remove `ollama.ts` file references
 - `.agents/plans/tool-calling-infrastructure.md` — remove Ollama mentions
 - `.agents/plans/mcp-integration-plan.md` — remove `ollama-section.tsx` references
@@ -665,10 +665,10 @@ After all priorities are complete:
 | `components/icons/README.md` | Remove OllamaIcon entry |
 | `.agents/context/architecture.md` | Remove Ollama from diagram |
 | `.agents/context/glossary.md` | Remove Ollama references |
-| `.agents/context/decisions/002-vercel-ai-sdk.md` | Remove Ollama provider |
-| `.agents/context/decisions/004-exa-vs-tavily.md` | Remove self-hoster references |
-| `.agents/context/research/competitive-feature-analysis.md` | Remove self-hosting mention |
-| `.agents/context/research/tool-calling-infrastructure.md` | Remove Ollama mentions |
+| `.agents/research/tech-stack-evaluation.md` | Remove Ollama provider |
+| `.agents/research/tool-calling-infrastructure.md` | Remove self-hoster references |
+| `.agents/research/competitive-feature-analysis.md` | Remove self-hosting mention |
+| `.agents/research/tool-calling-infrastructure.md` | Remove Ollama mentions |
 | `.agents/plans/thinking-reasoning-configuration.md` | Remove ollama.ts references |
 | `.agents/plans/tool-calling-infrastructure.md` | Remove Ollama mentions |
 | `.agents/plans/mcp-integration-plan.md` | Remove ollama-section.tsx references |

.agents/archive/installation-implementation-plan-2026-01.md

@@ -119,7 +119,7 @@
 # Check current status
 git status
 
-# Stage the evaluation document (historical path, now at .agents/context/research/)
+# Stage the evaluation document (historical path, now at .agents/research/)
 git add docs/youtube-transcript-evaluation.md
 
 # Review changes to layout
@@ -2001,7 +2001,7 @@ rm -rf app/api/youtube
 
 ### Project Documentation
 - [INSTALL.md](../../INSTALL.md) - Full installation guide
-- [YouTube Transcript Evaluation](../context/research/youtube-transcript-evaluation.md) - Detailed transcript implementation options
+- [YouTube Transcript Evaluation](../research/youtube-transcript-evaluation.md) - Detailed transcript implementation options
 
 ### UI Component Libraries
 - [shadcn/ui Docs](https://ui.shadcn.com/docs)

.agents/archive/model-registry-update-feb-2026.md

@@ -1,7 +1,7 @@
 # Model Registry Update — February 2026
 
 > **Created**: February 6, 2026
-> **Research**: `.agents/context/research/latest-models-february-2026.md`
+> **Research**: `.agents/research/latest-models-february-2026.md`
 > **Scope**: Deprecate retired models, update stale entries, add new generation models
 > **Skill**: Follow `@.agents/skills/add-model/SKILL.md` for the 3-step model addition pattern
 

.agents/context/ai-context-engineering-guide.md

@@ -1116,11 +1116,11 @@ The following areas require additional investigation or clarification:
 
 | Question | Resolution | Reference |
 |----------|------------|-----------|
-| Convex vs Supabase for AI-first apps? | **Convex** — Built-in RAG, vector search, TypeScript-first | `.agents/context/research/tech-stack-evaluation.md` |
-| Auth provider for Convex + payments? | **Clerk** — Native integration with both Convex and Flowglad | `.agents/context/research/tech-stack-evaluation.md` |
-| Payments platform? | **Flowglad** — Open-source, Clerk-native, developer-friendly | `.agents/context/research/tech-stack-evaluation.md` |
-| Testing strategy for AI apps? | **Critical paths only** — Auth, data transforms, rate limiting | `.agents/context/research/tech-stack-evaluation.md` |
-| Sub-agent architecture? | **4-agent design** — Transcript, Title, Thumbnail, Analytics | `.agents/context/research/tech-stack-evaluation.md` |
+| Convex vs Supabase for AI-first apps? | **Convex** — Built-in RAG, vector search, TypeScript-first | `.agents/research/tech-stack-evaluation.md` |
+| Auth provider for Convex + payments? | **Clerk** — Native integration with both Convex and Flowglad | `.agents/research/tech-stack-evaluation.md` |
+| Payments platform? | **Flowglad** — Open-source, Clerk-native, developer-friendly | `.agents/research/tech-stack-evaluation.md` |
+| Testing strategy for AI apps? | **Critical paths only** — Auth, data transforms, rate limiting | `.agents/research/tech-stack-evaluation.md` |
+| Sub-agent architecture? | **4-agent design** — Transcript, Title, Thumbnail, Analytics | `.agents/research/tech-stack-evaluation.md` |
 
 ### Remaining Questions
 
@@ -1215,8 +1215,8 @@ The following areas require additional investigation or clarification:
 - [Anthropic Model System Cards](https://www.anthropic.com/system-cards)
 
 ### Project-Specific Research (Not A Wrapper)
-- `.agents/context/research/tech-stack-evaluation.md` — Comprehensive tech stack evaluation (Convex vs Supabase, auth, payments)
-- `.agents/context/research/youtube-transcript-evaluation.md` — YouTube transcript extraction research
+- `.agents/research/tech-stack-evaluation.md` — Comprehensive tech stack evaluation (Convex vs Supabase, auth, payments)
+- `.agents/research/youtube-transcript-evaluation.md` — YouTube transcript extraction research
 - `AGENTS.md` — Finalized project configuration for AI agents
 
 ## Case Study: Not A Wrapper

.agents/context/architecture.md

@@ -290,4 +290,4 @@ components/           # Shadcn UI components (shared)
 
 ---
 
-*See `@.agents/context/research/tech-stack-evaluation.md` for detailed tech stack evaluation and decision rationale.*
+*See `@.agents/research/tech-stack-evaluation.md` for detailed tech stack evaluation and decision rationale.*