Merge v2611-followups: v2.6.11 apps/server close-hook caller + DiffPanel staging hint

CHANGELOG + roadmap (through v2.6.11) + openspec v2-6 Phase 3 fully closed (3.7 + apps/server close-hook caller done).

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

CHANGELOG.md

@@ -2,6 +2,10 @@
 
 All notable changes per release tag. Most recent on top, ordered by tag creation date (which matches the git history). Tag names follow `vMAJOR.MINOR.PATCH-slug` — the slug describes what shipped, so the tag name alone is enough to recall the batch.
 
+## v2.6.11-close-hooks-staging — 2026-06-01
+
+The two v2.6 follow-ups left after `v2.6.10-lifecycle-hardening`. **Server close-hook caller:** `apps/server` (BooChat) now fire-and-forgets BooCoder's Phase-3 close hooks so warm agent backends + worktrees tear down *immediately* on delete/archive instead of waiting for the idle-evict/reaper backstop — a new `coder-notify.ts` `notifyCoderClose(kind,id)` (reusing the v2.6.2 `BOOCODER_URL` reach, never-rejects) is `void`-called after the WS frame at session-delete (`POST /api/sessions/:id/close`) and chat archive / archive-all / delete (`POST /api/chats/:id/close`); an unreachable coder can never block or fail the user's delete/archive. **Staging-boundary hint (task 3.7):** the BooCoder DiffPanel now shows a muted one-liner when the selected provider can't see another agent's unapplied worktree edits — native boocode selected + external-agent-staged changes (or vice-versa) → "<agent>'s edits live in its worktree — BooCode won't see them until applied" — derived purely from the per-change `agent` + current provider, no new state. 6 new server tests (`coder-notify`), 537 server tests pass; web + server tsc/build clean. **With these the v2.6 openspec is fully closed** — only the live Smoke 2/2b/3 remain (manual exercise).
+
 ## v2.6.10-lifecycle-hardening — 2026-06-01
 
 v2.6 Phase 3 (the last phase) — lifecycle hardening of the warm-process backends. **Idle eviction + LRU cap:** the agent pool runs a 60s sweep that evicts backends/sessions idle past `AGENT_POOL_IDLE_TTL_MS` (30 min default) and any beyond `AGENT_POOL_MAX_LIVE` (10, LRU) — **never a busy one** (in-flight turn, double-checked via a new `isBusy()` backend hook); the worktree persists (DB-backed) and the next turn re-spawns + reattaches. The eviction/LRU/restart decisions are factored into a pure `lifecycle-decisions.ts` (modeled on the inference `selectPruneTargets` pattern). **Crash recovery:** lifts openchamber's health-monitor + busy-aware-restart + consecutive-failure + stale-busy-grace state machine into `opencode-server.ts` (with port reclaim) and `warm-acp.ts` — an opencode server crash settles in-flight turns as failed, marks the rows `crashed`, and recreates fresh sessions (a fresh server can't hold the old in-memory id), while a warm-ACP child crash re-`session/new`s next turn; the F.1 turn-guard and U.6 usage are preserved (their tests still pass). **Worktree reaper:** a periodic reaper removes orphan on-disk worktrees (no live `worktrees` row, 1h grace) behind a superset-style preflight that skips dirty/unpushed/unmerged work, with Paseo-style soft-delete (`status='archived'`). Plus close hooks (`/api/chats/:id/close`, `/api/sessions/:id/close`, awaiting the apps/server caller) and diff re-baseline after `apply_pending`. Built test-first — 35 new tests (`lifecycle-decisions` 22, `agent-pool` 13) + a DB-opt-in reconnect integration test; 215 coder tests pass, tsc + build clean. **This completes v2.6** (Phase 0–3 + F.1 + Phase 1-UX). Remaining follow-ups (out of v2.6 scope): the apps/server close-hook caller, the 3.7 DiffPanel staging-boundary hint (frontend), and live Smoke 2/2b/3.

apps/server/src/routes/chats.ts

@@ -4,6 +4,7 @@ import type { Sql } from '../db.js';
 import type { Broker } from '../services/broker.js';
 import type { Chat, Message } from '../types/api.js';
 import { getModelContext } from '../services/model-context.js';
+import { notifyCoderClose } from '../services/coder-notify.js';
 
 const CreateBody = z.object({
   name: z.string().min(1).max(200).optional(),
@@ -167,6 +168,9 @@ export function registerChatRoutes(
           chat_id: id,
           session_id: req.params.id,
         });
+        // Fire-and-forget per archived chat: tear down its warm agent backends
+        // on the coder. Best-effort — never blocks/fails the bulk archive.
+        void notifyCoderClose('chat', id, req.log);
       }
       return { archived: ids.length, ids };
     }
@@ -208,6 +212,9 @@ export function registerChatRoutes(
         chat_id: row.id,
         session_id: row.session_id,
       });
+      // Fire-and-forget: tear down this chat's warm agent backends + (last-chat)
+      // worktree on the coder. Best-effort — never blocks/fails the archive.
+      void notifyCoderClose('chat', row.id, req.log);
       reply.code(204);
       return null;
     }
@@ -248,6 +255,9 @@ export function registerChatRoutes(
         chat_id: row.id,
         session_id: row.session_id,
       });
+      // Fire-and-forget: tear down this chat's warm agent backends + (last-chat)
+      // worktree on the coder. Best-effort — never blocks/fails the delete.
+      void notifyCoderClose('chat', row.id, req.log);
       reply.code(204);
       return null;
     }

apps/server/src/routes/sessions.ts

@@ -5,6 +5,7 @@ import type { Config } from '../config.js';
 import type { Broker } from '../services/broker.js';
 import type { Session, WorktreeRiskReport } from '../types/api.js';
 import { getSetting } from './settings.js';
+import { notifyCoderClose } from '../services/coder-notify.js';
 
 const CreateBody = z.object({
   name: z.string().min(1).max(200).optional(),
@@ -513,6 +514,10 @@ export function registerSessionRoutes(
       }
       const project_id = deleted[0]!.project_id;
       broker.publishUserFrame('default', { type: 'session_deleted', session_id: id, project_id });
+      // Fire-and-forget: ask BooCoder to tear down this session's warm agent
+      // backends + worktree immediately. Best-effort — never blocks/fails the
+      // delete; the coder's idle-evict + orphan reaper backstop a missed call.
+      void notifyCoderClose('session', id, req.log);
       reply.code(204);
       return null;
     }

apps/server/src/services/__tests__/coder-notify.test.ts

@@ -0,0 +1,67 @@
+// v2.6.10 Phase 3 (server wiring) — notifyCoderClose fire-and-forget helper.
+//
+// The guarantee under test: the helper NEVER throws (so it can't break the
+// user's delete/archive path), targets the correct coder URL shape, and folds
+// every failure mode (non-2xx, network error) into a `false` result.
+
+import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest';
+import { notifyCoderClose } from '../coder-notify.js';
+
+const ORIGINAL_BOOCODER_URL = process.env.BOOCODER_URL;
+
+describe('notifyCoderClose', () => {
+  beforeEach(() => {
+    delete process.env.BOOCODER_URL;
+  });
+  afterEach(() => {
+    if (ORIGINAL_BOOCODER_URL === undefined) delete process.env.BOOCODER_URL;
+    else process.env.BOOCODER_URL = ORIGINAL_BOOCODER_URL;
+  });
+
+  it('POSTs the chat close hook at the default coder origin and resolves true on 2xx', async () => {
+    const fetcher = vi.fn().mockResolvedValue(new Response(null, { status: 200 }));
+    const ok = await notifyCoderClose('chat', 'chat-123', undefined, fetcher as unknown as typeof fetch);
+    expect(ok).toBe(true);
+    expect(fetcher).toHaveBeenCalledTimes(1);
+    const [url, init] = fetcher.mock.calls[0]!;
+    expect(url).toBe('http://boocoder:3000/api/chats/chat-123/close');
+    expect(init).toEqual({ method: 'POST' });
+  });
+
+  it('POSTs the session close hook with the sessions segment', async () => {
+    const fetcher = vi.fn().mockResolvedValue(new Response(null, { status: 200 }));
+    const ok = await notifyCoderClose('session', 'sess-abc', undefined, fetcher as unknown as typeof fetch);
+    expect(ok).toBe(true);
+    expect(fetcher.mock.calls[0]![0]).toBe('http://boocoder:3000/api/sessions/sess-abc/close');
+  });
+
+  it('honors BOOCODER_URL for the origin', async () => {
+    process.env.BOOCODER_URL = 'http://100.114.205.53:9502';
+    const fetcher = vi.fn().mockResolvedValue(new Response(null, { status: 200 }));
+    await notifyCoderClose('chat', 'c1', undefined, fetcher as unknown as typeof fetch);
+    expect(fetcher.mock.calls[0]![0]).toBe('http://100.114.205.53:9502/api/chats/c1/close');
+  });
+
+  it('resolves false on a non-2xx response (does not throw)', async () => {
+    const fetcher = vi.fn().mockResolvedValue(new Response(null, { status: 500 }));
+    const log = { debug: vi.fn() };
+    const ok = await notifyCoderClose('chat', 'c1', log, fetcher as unknown as typeof fetch);
+    expect(ok).toBe(false);
+    expect(log.debug).toHaveBeenCalledTimes(1);
+  });
+
+  it('resolves false on a network error (coder unreachable) — never rejects', async () => {
+    const fetcher = vi.fn().mockRejectedValue(new Error('ECONNREFUSED'));
+    const log = { debug: vi.fn() };
+    const ok = await notifyCoderClose('session', 's1', log, fetcher as unknown as typeof fetch);
+    expect(ok).toBe(false);
+    expect(log.debug).toHaveBeenCalledTimes(1);
+  });
+
+  it('does not require a logger', async () => {
+    const fetcher = vi.fn().mockRejectedValue(new Error('boom'));
+    await expect(
+      notifyCoderClose('chat', 'c1', undefined, fetcher as unknown as typeof fetch),
+    ).resolves.toBe(false);
+  });
+});

apps/server/src/services/coder-notify.ts

@@ -0,0 +1,64 @@
+// v2.6.10 Phase 3 (server wiring) — fire-and-forget BooCoder close hooks.
+//
+// BooCoder (apps/coder, host systemd) added close hooks in
+// apps/coder/src/routes/lifecycle.ts:
+//   POST /api/chats/:chatId/close      — evict the chat's warm (chat,agent)
+//                                        backends, close its opencode session,
+//                                        mark agent_sessions closed, and remove
+//                                        the shared worktree on the last chat.
+//   POST /api/sessions/:sessionId/close — loop the chat-close path for every
+//                                        chat in the session.
+//
+// apps/server (Docker) can't see the host worktree dirs or reach the warm agent
+// processes, so — exactly like the existing `worktree-risk` guard in
+// routes/sessions.ts — it signals the coder over HTTP and the coder does the
+// real teardown. This call is BEST-EFFORT: the coder's idle-pool eviction and
+// the orphan-worktree reaper backstop a missed/failed call. It MUST NEVER block
+// or fail the user's delete/archive — hence fire-and-forget with a swallowed
+// catch. We do not await the returned promise at the call sites.
+
+import type { FastifyBaseLogger } from 'fastify';
+
+export type CoderCloseKind = 'chat' | 'session';
+
+function coderOrigin(): string {
+  // Same env + default as routes/sessions.ts' worktree-risk fetch.
+  return process.env.BOOCODER_URL ?? 'http://boocoder:3000';
+}
+
+/**
+ * Fire-and-forget POST to the BooCoder close hook for a chat or session.
+ *
+ * Resolves to `true` if the coder acknowledged (HTTP 2xx), `false` otherwise
+ * (non-2xx or network error). Callers SHOULD NOT await this — invoke it and
+ * move on. The returned promise never rejects: every failure path is caught,
+ * logged at debug, and folded into a `false` result so an unreachable or
+ * erroring coder can't surface to the user's delete/archive request.
+ */
+export async function notifyCoderClose(
+  kind: CoderCloseKind,
+  id: string,
+  log?: Pick<FastifyBaseLogger, 'debug'>,
+  fetcher: typeof fetch = fetch,
+): Promise<boolean> {
+  const segment = kind === 'chat' ? 'chats' : 'sessions';
+  const url = `${coderOrigin()}/api/${segment}/${id}/close`;
+  try {
+    const res = await fetcher(url, { method: 'POST' });
+    if (!res.ok) {
+      log?.debug(
+        { kind, id, status: res.status },
+        'coder close hook returned non-2xx (best-effort; reaper backstops)',
+      );
+      return false;
+    }
+    log?.debug({ kind, id }, 'coder close hook acknowledged');
+    return true;
+  } catch (err) {
+    log?.debug(
+      { kind, id, err: err instanceof Error ? err.message : String(err) },
+      'coder close hook unreachable (best-effort; reaper backstops)',
+    );
+    return false;
+  }
+}

apps/web/src/components/panes/CoderPane.tsx

@@ -388,12 +388,14 @@ function usePendingChanges(sessionId: string) {
 function DiffPanel({
   changes,
   loading,
+  currentProvider,
   onRefresh,
   onApprove,
   onReject,
 }: {
   changes: PendingChange[];
   loading: boolean;
+  currentProvider: string;
   onRefresh: () => void;
   onApprove: (id: string) => void;
   onReject: (id: string) => void;
@@ -409,6 +411,29 @@ function DiffPanel({
       ? `Changes from ${distinctAgents.map((a) => providerLabel(a)).join(', ')}`
       : null;
 
+  // v2.6 §9c: staging-boundary caveat. External agents (opencode/goose/qwen/
+  // claude) edit *inside their worktree*; native boocode reads/writes the
+  // *project root* via pending_changes. Unapplied edits don't cross that
+  // boundary. When the currently-selected provider can't see another side's
+  // staged-but-unapplied edits, surface a muted one-liner. agent===null
+  // (manual) is boundary-neutral. Pure derivation — no new state/fetch.
+  const isNativeProvider = currentProvider === 'boocode';
+  const boundaryHint = (() => {
+    if (isNativeProvider) {
+      // Native boocode is selected: it won't see external-worktree edits.
+      const external = distinctAgents.filter((a) => a !== null && a !== 'boocode');
+      if (external.length === 0) return null;
+      const who =
+        external.length === 1
+          ? providerLabel(external[0]!)
+          : external.map((a) => providerLabel(a)).join(', ');
+      return `${who}'s edits live in its worktree — BooCode won't see them until applied.`;
+    }
+    // An external agent is selected: it won't see boocode's project-root edits.
+    if (!distinctAgents.includes('boocode')) return null;
+    return `BooCode's edits live in the project root — ${providerLabel(currentProvider)} won't see them until applied.`;
+  })();
+
   return (
     <div className="flex flex-col h-full border-t border-border">
       <div className="flex items-center justify-between px-3 py-1.5 border-b border-border bg-muted/30">
@@ -430,6 +455,14 @@ function DiffPanel({
           {mixedNote}
         </div>
       )}
+      {boundaryHint && (
+        <div
+          className="px-3 py-1 border-b border-border bg-muted/10 text-xs text-muted-foreground"
+          title={boundaryHint}
+        >
+          {boundaryHint}
+        </div>
+      )}
       <div className="flex-1 overflow-y-auto">
         {pending.length === 0 ? (
           <div className="flex items-center justify-center h-full text-sm text-muted-foreground">
@@ -914,6 +947,7 @@ export function CoderPane({
           <DiffPanel
             changes={changes}
             loading={loading}
+            currentProvider={agentConfig.provider}
             onRefresh={refresh}
             onApprove={approve}
             onReject={reject}

boocode_roadmap.md

@@ -348,7 +348,7 @@ Per-session Docker sandbox spawned by BooCoder on first write. Only project path
 
 -----
 
-## Shipped (v2.2.2–v2.6.10 — interactive ACP, provider lifecycle, persistent agent sessions, workspace UX)
+## Shipped (v2.2.2–v2.6.11 — interactive ACP, provider lifecycle, persistent agent sessions, workspace UX)
 
 All tags `vMAJOR.MINOR.PATCH-slug`, monotonic per minor, assigned at ship time (planning slugs differ — see the numbering-discipline note below). `CHANGELOG.md` is the canonical per-tag record. **Note on numbering divergence:** the *planned-feature* "v2.3 — Provider lifecycle" actually shipped under the **v2.5.4–v2.5.13** tags; the *planned-feature* "v2.4 — BooCoder as ACP agent" remains **unshipped** even though v2.4.0/v2.4.1 *tags* shipped unrelated content (Unsloth lifts, sidecar routing). The patch-tag thread and the conceptual-milestone thread have diverged — read tags as the ship record, the `## v2.x` feature sections below as the milestone plan. The v2.3.0–v2.5.1 tags were never CHANGELOG-backfilled; summarized here from commit bodies.
 
@@ -385,6 +385,7 @@ All tags `vMAJOR.MINOR.PATCH-slug`, monotonic per minor, assigned at ship time (
 - `v2.6.8-agent-attribution` — **v2.6 Phase 1-UX** (U.1–U.6), built by 3 parallel subagents over disjoint files. Backend: `pending_changes.agent` stamped at every queue site + flows through `listPending`; new `GET /api/sessions/:id/agent-sessions` route; opencode warm-server consumes `session.next.step.ended` → accumulates `input_tokens`/`output_tokens`/`cost` on `agent_sessions`. Frontend: DiffPanel per-row agent badges + multi-agent note; AgentComposerBar resumed/history/new-session chip (gated on optional `sessionId`, BooChat unaffected); shared `providerIcons.tsx` + `useAgentSessions` hook. 9 new tests; web+coder tsc clean. Both surfaces deployed (boocoder restart + `boocode` Docker rebuild). Phase 2/3 remain
 - `v2.6.9-warm-acp` — **v2.6 Phase 2:** goose/qwen run as **warm ACP backends** (one persistent `goose acp`/`qwen --acp` child + `ClientSideConnection` + ACP session per `(chat,agent)`, `initialize`+`session/new` once, reused across turns) instead of one-shot. New `WarmAcpBackend` (same `AgentBackend` interface as opencode); abort = `session/cancel` the prompt only (never kills the child); dispatcher routes goose/qwen chat-tab tasks via pure `shouldUseWarmBackend` (one-shot fallback kept for arena/MCP/`new_task`); `handleSessionUpdate` extracted to a shared pure `acp-event-map.ts` (one-shot path byte-identical). SDK concern resolved (`@agentclientprotocol/sdk@^0.22.1` has stable resume; moot warm, deferred to Phase 3). 15 new tests, 180 coder tests pass. Backend-only deploy (boocoder restart). **Smoke 2/2b pending live.** Phase 3 (lifecycle hardening) is the last v2.6 phase
 - `v2.6.10-lifecycle-hardening` — **v2.6 Phase 3 (final phase — completes v2.6).** Idle TTL eviction (`AGENT_POOL_IDLE_TTL_MS`=30min) + LRU cap (`AGENT_POOL_MAX_LIVE`=10), busy backends never evicted; pure `lifecycle-decisions.ts`. Crash recovery via openchamber's health-monitor + busy-aware-restart + stale-grace state machine in `opencode-server.ts` (+ port reclaim) + `warm-acp.ts` (opencode → fresh sessions; ACP → re-`session/new`; F.1 guard + U.6 usage preserved). Orphan worktree reaper (1h grace, superset-style dirty/unpushed preflight, Paseo soft-delete) + close hooks + re-baseline after apply. 35 new tests + DB-opt-in reconnect test; 215 coder tests pass. Backend-only deploy. **Follow-ups (out of v2.6 scope): apps/server close-hook caller, 3.7 DiffPanel staging hint (frontend), live Smoke 2/2b/3.** With this, **v2.6 persistent agent sessions is complete** (Phase 0–3 + F.1 + Phase 1-UX)
+- `v2.6.11-close-hooks-staging` — the two v2.6 follow-ups. **apps/server close-hook caller:** BooChat fire-and-forgets BooCoder's Phase-3 close hooks (new `coder-notify.ts`, never-rejects) on session-delete + chat archive/delete, so warm backends + worktrees tear down immediately (the idle-evict/reaper was the backstop). **Task 3.7 staging hint:** BooCoder DiffPanel shows a muted one-liner when the selected provider can't see another agent's unapplied worktree edits (pure derivation from per-change `agent` + current provider). 6 new server tests; web+server tsc/build clean; deploys via the `boocode` Docker container. **The v2.6 openspec is now fully closed** — only live Smoke 2/2b/3 remain (manual)
 
 -----
 

openspec/changes/v2-6-persistent-agent-sessions/tasks.md

@@ -54,17 +54,17 @@ ACP follows; hardening last.
       resumes the SAME `agent_session_id` (memory intact), boocode saw opencode's turns as
       history, all three shared the one worktree, and no agent was locked to the chat.
 
-## Phase 3 — Lifecycle hardening — ✅ SHIPPED `v2.6.10-lifecycle-hardening` (3.1–3.6; 3.7 frontend + apps/server close-hook caller are follow-ups)
+## Phase 3 — Lifecycle hardening — ✅ COMPLETE (`v2.6.10` 3.1–3.6; `v2.6.11` closed 3.7 + the apps/server close-hook caller)
 
 > **Lift (design §10):** hardening from **openchamber** (MIT, same warm-opencode-server architecture) — health-monitor + crash auto-restart + busy-aware restart + port reclaim (`killProcessOnPort`/`waitForPortRelease`) + stall-SSE = a concrete state machine for 3.1/3.2/3.6. Reaper (3.3/3.4): Paseo worktree-archive cascade + superset destroy-saga (preflight dirty/unpushed inspect) + LRU cap on warm-server Maps. Do crash-recovery + reaper together (shared supervision loop).
 
 - [x] 3.1 Idle TTL eviction per `(chat, agent)` (`AGENT_POOL_IDLE_TTL_MS`=30min) + LRU cap (`AGENT_POOL_MAX_LIVE`=10), busy never evicted; reattach next turn. Pure `lifecycle-decisions.ts` (TDD).
 - [x] 3.2 Crash recovery: openchamber health-monitor + busy-aware-restart + stale-grace state machine in `opencode-server.ts` (+ port reclaim) + `warm-acp.ts`. opencode → fresh sessions; ACP → re-`session/new`. F.1 guard + U.6 usage preserved.
-- [x] 3.3 Close hooks (`/api/chats/:id/close`, `/api/sessions/:id/close`) → `closeChat` evicts backends + archives the `worktrees` row + removes the worktree. *(apps/server caller is a follow-up; idle-evict + reaper backstop it.)*
+- [x] 3.3 Close hooks (`/api/chats/:id/close`, `/api/sessions/:id/close`) → `closeChat` evicts backends + archives the `worktrees` row + removes the worktree. **apps/server caller wired in `v2.6.11`** (`coder-notify.ts`, fire-and-forget on session-delete + chat archive/delete).
 - [x] 3.4 Orphan worktree reaper (periodic, 1h grace, superset-style dirty/unpushed preflight, Paseo soft-delete) + LRU cap on the pool.
 - [x] 3.5 Re-baseline `worktrees.base_commit` after a successful `apply_pending` (both apply routes).
 - [x] 3.6 Reconnect integration test (DB-opt-in): restart mid-session → next turn reattaches/recreates from `agent_sessions`/`worktrees`.
-- [ ] 3.7 Staging-boundary hint in DiffPanel (§9c) — **frontend follow-up** (apps/web; deferred — Sam has uncommitted web work).
+- [x] 3.7 Staging-boundary hint in DiffPanel (§9c) — `v2.6.11`: muted one-liner when the selected provider can't see another agent's unapplied worktree edits (derived from per-change `agent` + current provider; no new state).
 
 ## Tests — ⬜ REMAINING (none of T.1–T.3 exist yet)