aifirst: lift context/agents/tools runtime to lib-aifirst-ui

The mechanism (context surface registry, persona storage + hooks, tool
parser/dispatcher) is now generic and lives in @crema/aifirst-ui/{context,
agents,tools}. This template keeps only the arcadia-shaped configuration:

- agents.ts — owns DEFAULT_AGENTS + legacy/retired migration sets, calls
  configureAgents() at module load, re-exports the runtime
- admin-tools.ts — keeps the 19 arcadia tool definitions, binds the
  runtime via createToolRuntime(TOOLS), re-exports the bound functions
- admin-context.ts — deleted; 18 routes now import directly from
  @crema/aifirst-ui/context

Routes that import from ~/lib/agents and ~/lib/admin-tools are unchanged
(wrapper modules preserve the existing import surface).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

app/routes/assistant.tsx

@@ -90,7 +90,7 @@ wait_for assistant-ui-control
 \`\`\`"`
 
 
-import { formatAdminContextForPrompt } from "~/lib/admin-context"
+import { formatContextForPrompt } from "@crema/aifirst-ui/context"
 import {
   buildDenialMessages,
   classifyCalls,
@@ -111,7 +111,7 @@ function buildAdminPreface(activeAgent: Agent | undefined, uiControl: boolean):
   const persona = activeAgent
     ? `Active persona: ${activeAgent.name} — ${activeAgent.role}\n${activeAgent.prompt}`
     : ""
-  const ctx = formatAdminContextForPrompt()
+  const ctx = formatContextForPrompt()
   const parts = [
     "You are the operator's assistant inside Arcadia Admin. Be precise and direct. You have native function tools attached to this conversation — call them whenever the user asks about live platform state (counts, statuses, listings, lookups). Never invent tenant slugs, user counts, or statuses; if you need data, call a tool.",
     "Two retrieval surfaces exist for documentation/knowledge: `search_docs` (browser-side, BM25 over the bundled arcadia docs — fast, always available, small corpus) and `search_kb` (server-side, BM25 over arcadia-search — `docs` (arcadia parity), `operator-tools` (arcadia-search + arcadia-admin admin docs), `files` (uploaded files), plus any custom corpora the operator adds via /search). For questions about the bundled arcadia docs either is fine; prefer `search_kb` for richer hits or for content outside the bundled docs (uploaded files, the admin tooling itself, tenant-specific knowledge). If unsure what corpora exist, call `list_search_corpora`. When `search_kb` returns a chunk_id you want to expand, call `read_chunk(chunk_id, corpus)`. When the operator says results look stale or after they've uploaded new files, call `rebuild_search_corpus(tenant, corpus)`.",