feat: Reginald is real — model router + agent loop + query_project (P4)

The fixture chat panel is now a working agent. Ask Reginald a question and it
consults the real project through a tool loop, then answers in grounded prose.
Read-only v0 — writes still go through the propose-approve controls.

core (@commitea/core/agent):
- chat-client: OpenAI-wire chat completions over an injected fetch (same seam as
  gitea). Points at any OpenAI-compatible endpoint (LM Studio/Ollama/OpenAI).
- model-router: small model for prose + the read tool; big model reserved for
  later decomposition (pickModel).
- agent-loop: runAgentTurn drives call→tool→result→call until prose (or a step
  budget), recording each tool step. Injected complete + execute → fully testable.
- query-project: the single read tool's engine — compact focus/board/calibration/
  issue/search views built from scheduler + lifecycle + calibration; unbuilt views
  return a notImplemented marker (never fabricated). The model reports, never computes.
- agent-tools: query_project declaration + Reginald's system prompt.

app:
- main model bridge (model:status, model:chat) runs the loop; query_project
  reconciles the repo and builds the view. Model traffic stays in main (token/CSP).
  gitea.ts refactored to share getGiteaClient + reconcileSnapshot.
- preload + global.d.ts expose the model bridge; useChat drives the panel — real
  agent turn when a model is configured, scripted fixture reply otherwise (so
  fixture e2e is unchanged). A subtle "consulted the project" activity line.

Model config (env, defaults to LM Studio on :1234): COMMITEA_MODEL_URL /
_SMALL (google/gemma-4-e4b) / _BIG (qwen/qwen3.6-35b-a3b). COMMITEA_E2E=1 keeps
it unconfigured so the panel stays scripted.

Verified: 88 core tests green (14 agent: client parse, loop tool/error/budget,
all views) + a gated live integration test. Desktop typecheck clean, 14 fixture
e2e green. Gated live e2e drives the real app against gitea + gemma-4-e4b: asked
"what now?", Reginald called query_project and answered "focus is on issue #2"
(the real scheduler pick).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

apps/desktop/src/main/gitea.ts

@@ -11,11 +11,13 @@ import { dirname, join } from 'node:path'
 
 import {
   createGiteaClient,
+  type GiteaClient,
   type GiteaConfig,
   type GiteaLabel,
   type IssueChange,
   type LifecycleEvent,
   planIssueChange,
+  type ProjectSnapshot,
 } from '@commitea/core'
 import { ipcMain } from 'electron'
 
@@ -50,28 +52,45 @@ function resolveConfig(): GiteaConfig | null {
   }
 }
 
-export function registerGiteaIpc(): void {
-  const config = resolveConfig()
-  const client = config ? createGiteaClient(config, fetch) : null
-  const repo = config ? `${config.owner}/${config.repo}` : null
+// Memoized client so both the gitea and model bridges share one instance.
+let sharedClient: GiteaClient | null | undefined
+export function getGiteaClient(): GiteaClient | null {
+  if (sharedClient === undefined) {
+    const config = resolveConfig()
+    sharedClient = config ? createGiteaClient(config, fetch) : null
+  }
+  return sharedClient
+}
 
-  ipcMain.handle('gitea:status', () => ({ configured: !!config, repo }))
+/** Full reconcile: issues + milestones + native deps + lifecycle timelines. */
+export async function reconcileSnapshot(
+  client: GiteaClient,
+): Promise<ProjectSnapshot & { milestones: Awaited<ReturnType<GiteaClient['listMilestones']>> }> {
+  const [issues, milestones] = await Promise.all([client.listIssues(), client.listMilestones()])
+  // dependency edges among the open scope (the scheduler only plans what's left)
+  const open = issues.filter((i) => i.state === 'open')
+  const perIssue = await Promise.all(
+    open.map(async (i) => ({ issue: i.number, dependsOn: await client.getIssueDependencies(i.number) })),
+  )
+  const deps = perIssue.flatMap(({ issue, dependsOn }) => dependsOn.map((d) => ({ issue, dependsOn: d })))
+  // lifecycle timelines for every issue (open → columns/badges, closed → calibration actuals)
+  const timelineEntries = await Promise.all(
+    issues.map(async (i) => [i.number, await client.getIssueTimeline(i.number)] as const),
+  )
+  const timelines: Record<number, LifecycleEvent[]> = Object.fromEntries(timelineEntries)
+  return { issues, milestones, deps, timelines }
+}
+
+export function registerGiteaIpc(): void {
+  const client = getGiteaClient()
+  const repo = client ? `${process.env.GITEA_OWNER ?? 'christian'}/${process.env.GITEA_REPO ?? 'commitea'}` : null
+
+  ipcMain.handle('gitea:status', () => ({ configured: !!client, repo }))
 
   ipcMain.handle('gitea:reconcile', async () => {
     if (!client) return { configured: false, issues: [], milestones: [], deps: [], timelines: {} }
-    const [issues, milestones] = await Promise.all([client.listIssues(), client.listMilestones()])
-    // dependency edges among the open scope (the scheduler only plans what's left)
-    const open = issues.filter((i) => i.state === 'open')
-    const perIssue = await Promise.all(
-      open.map(async (i) => ({ issue: i.number, dependsOn: await client.getIssueDependencies(i.number) })),
-    )
-    const deps = perIssue.flatMap(({ issue, dependsOn }) => dependsOn.map((d) => ({ issue, dependsOn: d })))
-    // lifecycle timelines for every issue (open → columns/badges, closed → calibration actuals)
-    const timelineEntries = await Promise.all(
-      issues.map(async (i) => [i.number, await client.getIssueTimeline(i.number)] as const),
-    )
-    const timelines: Record<number, LifecycleEvent[]> = Object.fromEntries(timelineEntries)
-    return { configured: true, issues, milestones, deps, timelines }
+    const snap = await reconcileSnapshot(client)
+    return { configured: true, ...snap }
   })
 
   ipcMain.handle('gitea:getIssue', async (_event, index: number) => {