Compare commits

...

11 Commits

Author SHA1 Message Date
nntrivi2001 c28395402b chore: v2.8.40 — a Codex account switch reaches running sessions 2026-10-29 15:59:19 +07:00
nntrivi2001 7798530528 fix(profiles): a Codex switch reaches running and new sessions
Codex >= 0.164 sends every request through a shared app-server daemon that
outlives its TUIs and keeps the login it started with, so writing a new
auth.json changed nothing: open sessions and even new ones stayed on the
old (out-of-quota) account. After a switch, restart that CODEX_HOME's
daemon when one runs; open TUIs reconnect to it on their own (verified on
0.164.0 with two keys). The response says so (following), and the toast
tells the user running sessions follow.
2026-10-29 15:54:20 +07:00
nntrivi2001 25f63d9ad3 fix(terminal): rebuild the WebGL glyph atlas when the page comes back
A browser can drop the atlas's GPU memory without a context-loss event
(sleep/resume, a long-backgrounded tab, display/zoom change on macOS),
leaving every cell painted with a clipped or foreign glyph.
2026-10-29 15:44:16 +07:00
nntrivi2001 d060d288fb fix(sessions): Top tools counts each tool call once
The session analysis page's Top tools list counted every event carrying a
tool_name, so each call was counted twice (its PreToolUse and its
PostToolUse) while the Tool calls total beside it counted once — a session
with 840 calls showed Bash at 1,533. The list now counts the same way as
the total: PreToolUse, or PostToolUse in a session with none (imported).
2026-10-29 14:53:45 +07:00
nntrivi2001 6099841cef fix(sessions): archiving a live session stops its pane
Archive on a session a live pane held only hid its row, and the list
deliberately does not bring such a tab back, so the pane kept running
with nothing on screen showing it — a member archived five sessions
believing them closed and kept ~2.7 GB of claude/omp/postgres-mcp alive
for up to two days. Archive on a live session is now a Stop (endRuns,
the tab X's confirm, DELETE /api/run/:id?archive=1) in the sidebar row,
the workspace home's Recents and their right-click menu.
2026-10-29 14:26:25 +07:00
nntrivi2001 d70a7c9403 chore: v2.8.39 — no one can take a colleague's display name 2026-10-29 13:47:31 +07:00
nntrivi2001 63eaa739d4 fix(accounts): a display name may not read as another account's
A member kept renaming themselves to a colleague's display name, plus
"(you)", so the CPU/RAM card and every owner list showed two of that person.
A display name is now refused (409 ENAMETAKEN) when, with marks, case,
spaces, punctuation, bracketed parts and look-alike Cyrillic/Greek letters
ignored, it equals another account's display name or username — in
Settings → Account and on the Users page alike. Keeping the name one already
has stays allowed, so an older clash does not block an admin's other edits.
2026-10-29 13:47:20 +07:00
nntrivi2001 9f8dc9082e chore: v2.8.38 — every user row shows its avatar 2026-10-29 13:40:39 +07:00
nntrivi2001 f1a90e120e fix(users): every user row shows its avatar, not only its initials
System Health's per-user CPU/RAM card asked UserAvatar for the viewer's own
picture with no user id, and the host stats never carried avatarAt, so every
card fell back to letters. The groups now carry avatarAt and the card passes
userId. The user editor's header, the pool board's member chips and the pool
member picker drew hand-made initials; they use UserAvatar too.
2026-10-29 13:40:28 +07:00
nntrivi2001 db5a7776cc chore: v2.8.37 — the Artifacts installs list is never cut off 2026-10-29 11:25:10 +07:00
nntrivi2001 6c2c094f64 fix(artifacts): the installs list floats over the page, never cut off
It was an absolute box inside the card list, whose overflow-hidden clipped it
at the last card, and it opened to the right off a button near the edge. It
is now a portal at the button: below it, above when there is no room, kept
inside the window, following scroll and resize.
2026-10-29 11:25:10 +07:00
45 changed files with 3648 additions and 144 deletions
+6 -5
View File
@@ -1,4 +1,8 @@
{
"env": {
"ANTHROPIC_BASE_URL": "",
"ANTHROPIC_AUTH_TOKEN": ""
},
"permissions": {
"allow": [
"Bash(npm run:*)",
@@ -36,11 +40,8 @@
"Bash(node -e ':*)",
"Bash(node -e \"require\\('./server/lib/transcript-cache'\\); require\\('./scripts/import-history'\\); console.log\\('All modules load OK'\\)\")",
"Bash(claude --version)",
"Bash(gh pr *)"
"Bash(gh pr *)",
"Bash(grep *)"
]
},
"env": {
"ANTHROPIC_BASE_URL": "",
"ANTHROPIC_AUTH_TOKEN": ""
}
}
+1 -1
View File
@@ -1,6 +1,6 @@
# Claude Code Working Guide
_Last synced: v2.8.36 (2026-10-29) — an HTML page opens in a modal over the screen (Changes Preview, Chat)._ A change that makes any statement in this file, `AGENTS.md`, or a README wrong — a moved path, a new or changed extension point, a retired rule — updates that file in the same change-set and bumps this marker. **Bump = replace the version/date and a ≤ 1-line summary; never append a changelog here** — details belong in the commit message and `docs/`, because this whole file is loaded into every session.
_Last synced: v2.8.40 (2026-10-29) — a Codex profile switch restarts Codex's app-server daemon so sessions follow._ A change that makes any statement in this file, `AGENTS.md`, or a README wrong — a moved path, a new or changed extension point, a retired rule — updates that file in the same change-set and bumps this marker. **Bump = replace the version/date and a ≤ 1-line summary; never append a changelog here** — details belong in the commit message and `docs/`, because this whole file is loaded into every session.
## Project mission
- Maintain a reliable local-first dashboard for Claude Code session monitoring.
File diff suppressed because one or more lines are too long
File diff suppressed because it is too large Load Diff
@@ -3,7 +3,8 @@
* @description Tests for the sidebar's workspace list: one flat row per
* workspace in the stored order, a row expanding to its sessions (remembered),
* a live session opening its pane and an ended one its transcript, the
* workspace's OWN sessions only, a session's state dot, rename and archive,
* workspace's OWN sessions only, a session's state dot, rename and archive
* (a live one's pane stopped first),
* selection
* by URL, Alt+arrow reordering kept in localStorage, double-click rename, the
* search, the row menu's items by kind and tool, the status dot, and the
@@ -225,6 +226,8 @@ describe("WorkspaceTree", () => {
await user.type(input, "Better name{Enter}");
expect(api.sessions.patch).toHaveBeenCalledWith("s-old", { name: "Better name" });
await user.click(screen.getByTestId("workspace-session-archive-s-live"));
// A live session's Archive stops its pane first — never a hidden pane left running.
expect(api.run.kill).toHaveBeenCalledWith("agenthub-lane-1-2", { archive: true });
expect(api.sessions.patch).toHaveBeenCalledWith("s-live", { archived: true });
});
+6 -1
View File
@@ -683,7 +683,12 @@ export function HostSections({ host, heap }: { host: HostState; heap?: HeapInfo
aria-expanded={isOpen}
className="flex w-full items-center gap-2 text-left disabled:cursor-default"
>
<UserAvatar name={name} className="h-6 w-6 text-[10px]" />
<UserAvatar
name={name}
userId={g.userId}
avatarAt={g.avatarAt}
className="h-6 w-6 text-[10px]"
/>
<span className="min-w-0 flex-1 truncate text-xs text-fg-secondary">
{name}
</span>
@@ -69,6 +69,7 @@ const stats: HostStats = {
const api = vi.hoisted(() => ({
host: { stats: vi.fn(), history: vi.fn() },
me: { avatarUrl: () => null },
users: { avatarUrl: () => null },
}));
vi.mock("../../../lib/api", async (importOriginal) => {
const actual = await importOriginal<Record<string, unknown>>();
+17 -16
View File
@@ -14,7 +14,7 @@
* a catalog tool's bypass — a YOLO chip for Codex) and Local / New worktree — the workspace's Recents — the sessions it
* owns and has not archived (`?lane=&archived=0`, with each one's state dot,
* **Show more** past the first page, and the sidebar row's Rename/Archive in
* its right-click menu), then the archived ones folded with Unarchive; a live session opens its pane,
* its right-click menu — Archive on a live session stops its pane first, `endRuns`), then the archived ones folded with Unarchive; a live session opens its pane,
* any other its conversation (`&session=`, `SessionReplay`), nothing is resumed — and on the right its context: Instructions, Memory (read-only), Context files and
* Folders (`WorkspaceContext.tsx`, also folded at the bottom of the Changes panel).
* @author Nguyễn Ngọc Trí Vĩ <vinnt@smartgift.vn>
@@ -22,6 +22,7 @@
import { isSubmitEnter } from "../../lib/keys";
import { confirmDialog } from "../../lib/confirm";
import { endRuns } from "../../lib/stopSession";
import { useEffect, useRef, useState } from "react";
import { useTranslation } from "react-i18next";
import { useNavigate } from "react-router-dom";
@@ -307,30 +308,30 @@ function HomeBody({
}
};
const archiveSession = async (s: Session) => {
const live = !!liveRunFor(lane, liveRuns, s.id);
const run = liveRunFor(lane, liveRuns, s.id);
// A live one is stopped first, as the tab's X does (its own confirm):
// archiving only its row left a pane running that no list showed.
if (run) {
if ((await endRuns([run], { name: s.name || t("tree.newSession"), t })).length)
await patchSession(s, { archived: true });
return;
}
const wt = sessionWorktreeName(lane, s);
const ok = await confirmDialog(
wt && !live
wt
? {
title: t("tree.archiveWorktreeTitle"),
message: t("tree.archiveWorktreeBody", { name: wt }),
confirmLabel: t("tree.archiveWorktreeConfirm"),
}
: live
? {
title: t("tree.archiveLiveTitle"),
message: t("tree.archiveLiveBody"),
confirmLabel: t("sessionMenu.archive"),
destructive: false,
}
: {
title: t("tree.archiveSession"),
confirmLabel: t("sessionMenu.archive"),
destructive: false,
}
: {
title: t("tree.archiveSession"),
confirmLabel: t("sessionMenu.archive"),
destructive: false,
}
);
if (!ok) return;
if (wt && !live) {
if (wt) {
try {
await api.lanes.removeSessionWorktree(lane.id, s.id).catch((err: unknown) => {
if ((err as { code?: string }).code !== "ENOWORKTREE") throw err;
@@ -24,7 +24,7 @@
* The list shows `SESSION_LIMIT` rows and a **See all (N more)** under them that
* loads every one, then a **Show less** that goes back to `SESSION_LIMIT`. A
* live tab whose session was archived is not brought back as a New session,
* and archiving a live one asks first. F2 renames the focused row too; each
* and archiving a live one STOPS it first (`endRuns`, the tab X's confirm). F2 renames the focused row too; each
* dot carries its meaning as screen-reader text. `readOnly` (another user's
* workspace, for an analyst): no live tabs, no rename/archive/menu, and a row
* opens the session's analysis page (`/sessions/<id>`) — never a pane or a resume.
@@ -32,7 +32,7 @@
*/
import { useEffect, useState, type ReactNode } from "react";
import { confirmDialog } from "../../lib/confirm";
import { endRuns } from "../../lib/stopSession";
import { isSubmitEnter } from "../../lib/keys";
import { Archive, GitBranch, Pin, PinOff, SquareTerminal } from "lucide-react";
import { useTranslation } from "react-i18next";
@@ -363,17 +363,10 @@ export function WorkspaceSessionList({
};
const archive = async (s: Session) => {
if (sessionWorktreeName(lane, s) && !runFor(s)) return setArchiving(s);
// A live one keeps running in its tab; archiving only hides its row.
if (
runFor(s) &&
!(await confirmDialog({
title: t("tree.archiveLiveTitle"),
message: t("tree.archiveLiveBody"),
confirmLabel: t("sessionMenu.archive"),
destructive: false,
}))
)
return;
// A live one is stopped first, as the tab's X does: archiving only its row
// left a pane running that no list showed any more.
const run = runFor(s);
if (run && !(await endRuns([run], { name: s.name || t("tree.newSession"), t })).length) return;
await patch(s, { archived: true });
};
const archiveWithWorktree = async (s: Session) => {
+14 -1
View File
@@ -468,7 +468,8 @@ export const TerminalView = memo(function TerminalView({
// A terminal mounted on a hidden tab sits in a 0×0 box (`display:none`):
// fitting it there shrinks xterm to its 12×6 minimum, and that size would
// reach tmux with the attach. It stays unfitted until it is shown.
const hasBox = () => !!containerRef.current?.clientWidth && !!containerRef.current.clientHeight;
const hasBox = () =>
!!containerRef.current?.clientWidth && !!containerRef.current.clientHeight;
if (hasBox()) fit.fit();
// Take the keyboard as soon as the pane is on screen: the terminal IS the
// page here, and having to click it first meant the first thing typed went
@@ -761,6 +762,16 @@ export const TerminalView = memo(function TerminalView({
refit();
};
document.fonts?.addEventListener("loadingdone", onFontLoaded);
// The browser may drop or reuse the WebGL glyph atlas's GPU memory without
// a context-loss event (sleep/resume, a long-backgrounded tab, a display or
// zoom change on macOS): xterm keeps slicing the old coordinates and every
// cell shows a clipped or foreign glyph while the buffer stays correct.
// Rebuilding the atlas when the page comes back is cheap and redraws clean.
const onPageShown = () => {
if (document.visibilityState === "visible") term.clearTextureAtlas();
};
document.addEventListener("visibilitychange", onPageShown);
window.addEventListener("focus", onPageShown);
// Settings → Appearance changed the terminal's font or size: an open
// pane takes it now, re-measured and re-fitted like a font landing.
const onFontPref = () => {
@@ -779,6 +790,8 @@ export const TerminalView = memo(function TerminalView({
searchRef.current = null;
resizeObserver.disconnect();
document.fonts?.removeEventListener("loadingdone", onFontLoaded);
document.removeEventListener("visibilitychange", onPageShown);
window.removeEventListener("focus", onPageShown);
window.removeEventListener(FONT_SIZE_EVENT, onFontPref);
term.textarea?.removeEventListener("focus", reassert);
detachIme();
@@ -262,11 +262,11 @@ function LabelEdit({
function useRunner(load: () => Promise<unknown>) {
const [busy, setBusy] = useState<string | null>(null);
const run = useCallback(
async (key: string, fn: () => Promise<unknown>, success: string) => {
async <T,>(key: string, fn: () => Promise<T>, success: string | ((r: T) => string)) => {
setBusy(key);
try {
await fn();
toast.success(success);
const r = await fn();
toast.success(typeof success === "function" ? success(r) : success);
await load();
} catch (err: unknown) {
toast.error(err);
@@ -399,7 +399,14 @@ function OtherProfiles({
void run(
`switch:${p.name}`,
() => api.toolProfiles.activate(tool, p.name),
t("toolProfiles.switchSuccess", { name: p.label, tool: label })
(r) =>
t(
r.following ? "toolProfiles.switchFollowing" : "toolProfiles.switchSuccess",
{
name: p.label,
tool: label,
}
)
)
}
className={btnCls}
@@ -800,9 +807,7 @@ function ProfileUsageLine({
/** Org / seat / plan line shared by the active card and each profile row. */
function IdentityLine({ identity }: { identity: CcAccountIdentity }) {
const parts = [identity.organization, identity.plan || identity.subscriptionType].filter(
Boolean
);
const parts = [identity.organization, identity.plan || identity.subscriptionType].filter(Boolean);
if (parts.length === 0) return null;
return <div className="text-[10px] text-fg-muted mt-0.5">{parts.join(" · ")}</div>;
}
+9 -1
View File
@@ -13,6 +13,7 @@ import { useTranslation } from "react-i18next";
import { Search } from "lucide-react";
import type { AuthUser, ServerUserUsage } from "../../lib/api";
import { Checkbox } from "../Checkbox";
import { UserAvatar } from "./UserAvatar";
export const ROLES: AuthUser["role"][] = ["admin", "developer", "analyst"];
/** A role, every user, or the users in no pool group. */
@@ -168,7 +169,14 @@ export function MemberPicker({
checked={chosen.has(u.id)}
onChange={(on) => setMany([u.id], on)}
label={
<span>
<span className="inline-flex items-center gap-1.5">
<UserAvatar
name={u.displayName || u.username}
userId={u.id}
avatarAt={u.avatarAt}
letters={1}
className="h-5 w-5 text-[10px]"
/>
{u.username}
{u.disabled && <span className="ml-1 text-fg-muted">({t("disabled")})</span>}
{note?.(u)}
+8 -6
View File
@@ -36,6 +36,7 @@ import { toast } from "../../lib/toast";
import { ConfirmModal } from "../ConfirmModal";
import { AiToolIcon } from "../AiToolIcon";
import { MemberPicker } from "./MemberPicker";
import { UserAvatar } from "./UserAvatar";
/** Pool tools with their labels, in display order. */
export const POOL_TOOL_LABELS: { id: PoolTool; label: string }[] = [
@@ -363,12 +364,13 @@ export function PoolGroups({
} ${isDragged(drag) ? "opacity-40" : ""}`}
data-testid="pool-member"
>
<span
className="flex h-4 w-4 items-center justify-center rounded-full bg-accent/15 text-[9px] font-semibold uppercase text-accent"
aria-hidden
>
{name.charAt(0)}
</span>
<UserAvatar
name={name}
userId={id}
avatarAt={u?.avatarAt}
letters={1}
className="h-4 w-4 text-[9px]"
/>
{name}
{u && <span className="text-[10px] text-fg-muted">{t(`roles.${u.role}`)}</span>}
{isAdmin && from && (
+3 -1
View File
@@ -4,7 +4,9 @@
* (`GET /api/me/avatar`) or, with `userId`, another's (`GET
* /api/users/:id/avatar`, staff only), versioned by `avatarAt`; without one —
* or when the image fails to load — the first letters of its name. Used by the
* sidebar's account row, its workspace owner groups and Settings → Account.
* sidebar's account row, its workspace owner groups, Settings → Account, the
* Users page and its editor, the pool board and member picker, and System
* Health's per-user CPU/RAM card.
* @author Nguyễn Ngọc Trí Vĩ <vinnt@smartgift.vn>
*/
@@ -33,6 +33,7 @@ import { copyText } from "../../lib/clipboard";
import { isSubmitEnter } from "../../lib/keys";
import { passwordProblems } from "../../lib/passwordRules";
import { PasswordRules } from "./PasswordRules";
import { UserAvatar } from "./UserAvatar";
import { useDialog } from "../../hooks/useDialog";
import { Switch } from "../Switch";
import { Checkbox } from "../Checkbox";
@@ -243,12 +244,12 @@ export function UserEditDialog({
className="flex max-h-[90vh] w-full max-w-2xl flex-col overflow-hidden rounded-xl border border-border bg-surface-1 shadow-xl"
>
<div className="flex items-start gap-3 border-b border-border px-5 py-4">
<span
aria-hidden
className="flex h-10 w-10 flex-shrink-0 items-center justify-center rounded-full bg-accent/15 text-sm font-semibold uppercase text-accent"
>
{shown.slice(0, 2)}
</span>
<UserAvatar
name={shown}
userId={user.id}
avatarAt={user.avatarAt}
className="h-10 w-10 text-sm"
/>
<div className="min-w-0 flex-1">
<h2 id="user-edit-title" className="text-base font-semibold text-fg-primary">
{t("edit.title", { name: user.displayName || user.username })}
@@ -33,7 +33,7 @@ const accountPool = vi.hoisted(() => ({
updateGroup: vi.fn(),
deleteGroup: vi.fn(),
}));
const users = vi.hoisted(() => ({ list: vi.fn() }));
const users = vi.hoisted(() => ({ list: vi.fn(), avatarUrl: () => null }));
const aiTools = vi.hoisted(() => ({
list: vi.fn(),
ompSharing: vi.fn(),
@@ -13,7 +13,12 @@ import { render, screen, waitFor } from "@testing-library/react";
import userEvent from "@testing-library/user-event";
import type { ServerUserUsage } from "../../../lib/api";
const users = vi.hoisted(() => ({ get: vi.fn(), update: vi.fn(), remove: vi.fn() }));
const users = vi.hoisted(() => ({
get: vi.fn(),
update: vi.fn(),
remove: vi.fn(),
avatarUrl: () => null,
}));
const confirmDialog = vi.hoisted(() => vi.fn());
vi.mock("../../../lib/confirm", () => ({ confirmDialog }));
const lanes = vi.hoisted(() => ({ list: vi.fn() }));
@@ -16,6 +16,7 @@ const users = vi.hoisted(() => ({
setQuotaDefault: vi.fn(),
update: vi.fn(),
get: vi.fn(),
avatarUrl: () => null,
}));
const lanes = vi.hoisted(() => ({ list: vi.fn() }));
vi.mock("../../../lib/api", async (importOriginal) => {
-2
View File
@@ -505,8 +505,6 @@
"changes.loadMore": "Load more commits",
"changes.commentPasted": "Pasted into tab {{tab}} — press Enter there to send it",
"changes.splitter": "Resize the changes panel",
"tree.archiveLiveTitle": "Archive a live session?",
"tree.archiveLiveBody": "Its tab keeps running; archiving only hides it from this list. Unarchive it from the archived fold.",
"tree.state.awaiting": "waiting for you",
"tree.state.error": "error",
"tree.state.live": "live",
+1
View File
@@ -1396,6 +1396,7 @@
"noActive": "{{tool}} holds no login on this machine.",
"noProfiles": "Could not load the profiles.",
"switchSuccess": "Switched {{tool}} to \"{{name}}\". Start a new session to use it.",
"switchFollowing": "Switched {{tool}} to \"{{name}}\" — running sessions follow it.",
"deleteConfirmBody": "This removes the saved {{tool}} login. To use it again, sign in from a terminal and save it once more.",
"loginHint": "Add another {{tool}} login by signing in from a terminal first:",
"subtitle": {
+1
View File
@@ -79,6 +79,7 @@
"EBADUSER": "Username: letters, digits, '.', '_' or '-' (up to 64).",
"EBADPASSWORD": "Password cannot be empty.",
"EBADEMAIL": "That email address does not look valid.",
"ENAMETAKEN": "Another account already uses that name. Pick one that is yours.",
"EBADFIELD": "One of the values is not valid.",
"ENOLANE": "That workspace no longer exists.",
"ENOTOWNED": "That workspace is not this user's any more.",
-2
View File
@@ -505,8 +505,6 @@
"changes.loadMore": "Tải thêm commit",
"changes.commentPasted": "Đã dán vào tab {{tab}} — nhấn Enter ở đó để gửi",
"changes.splitter": "Đổi độ rộng khung thay đổi",
"tree.archiveLiveTitle": "Lưu trữ một session đang chạy?",
"tree.archiveLiveBody": "Tab của nó vẫn chạy; lưu trữ chỉ ẩn nó khỏi danh sách này. Bỏ lưu trữ trong mục đã lưu trữ.",
"tree.state.awaiting": "đang chờ bạn",
"tree.state.error": "lỗi",
"tree.state.live": "đang chạy",
+1
View File
@@ -1384,6 +1384,7 @@
"noActive": "{{tool}} chưa có đăng nhập nào trên máy này.",
"noProfiles": "Không tải được danh sách hồ sơ.",
"switchSuccess": "Đã chuyển {{tool}} sang \"{{name}}\". Mở phiên mới để dùng.",
"switchFollowing": "Đã chuyển {{tool}} sang \"{{name}}\" — các phiên đang chạy theo luôn.",
"deleteConfirmBody": "Thao tác này xoá đăng nhập {{tool}} đã lưu. Muốn dùng lại, hãy đăng nhập từ terminal và lưu lại lần nữa.",
"loginHint": "Thêm đăng nhập {{tool}} khác bằng cách đăng nhập từ terminal trước:",
"subtitle": {
+1
View File
@@ -79,6 +79,7 @@
"EBADUSER": "Tên đăng nhập: chữ, số, '.', '_' hoặc '-' (tối đa 64 ký tự).",
"EBADPASSWORD": "Mật khẩu không được để trống.",
"EBADEMAIL": "Địa chỉ email không hợp lệ.",
"ENAMETAKEN": "Tên này đã có tài khoản khác dùng. Hãy chọn tên của chính bạn.",
"EBADFIELD": "Có một giá trị không hợp lệ.",
"ENOLANE": "Workspace đó không còn tồn tại.",
"ENOTOWNED": "Workspace đó không còn thuộc người này.",
+3 -1
View File
@@ -684,6 +684,8 @@ export interface HostUsageGroup {
username?: string;
/** The user's display name, null when none is set (absent from an older server). */
displayName?: string | null;
/** When the user's avatar was last set (null = none; absent from an older server). */
avatarAt?: string | null;
/** % of the whole machine's CPU (100 = every core busy) over the sample. */
cpu: number;
/** Resident memory, bytes (shared pages count in every process). */
@@ -1399,7 +1401,7 @@ export const api = {
),
/** POST /api/tool-profiles/:tool/activate — the outgoing login is saved back first. */
activate: (tool: string, name: string) =>
request<{ ok: true; name: string; savedOutgoing: string | null }>(
request<{ ok: true; name: string; savedOutgoing: string | null; following?: boolean }>(
`/tool-profiles/${encodeURIComponent(tool)}/activate`,
{ method: "POST", body: JSON.stringify({ name }) }
),
+2 -1
View File
@@ -6,7 +6,8 @@
* message. Both ask first when any pane holds an AI agent (it may be
* mid-turn); a plain shell holds no session, so it neither asks nor sleeps.
* Used by the tab bar (`ConsoleTabs`), the split panes (`Workspace`) and a
* session's right-click menu (`SessionMenu`).
* session's right-click menu (`SessionMenu`), and a live session's Archive
* (`WorkspaceSessionList`, `WorkspaceHome`) — archiving never leaves a pane running.
* @author Nguyễn Ngọc Trí Vĩ <vinnt@smartgift.vn>
*/
+97 -62
View File
@@ -48,7 +48,8 @@
* @author Nguyễn Ngọc Trí Vĩ <vinnt@smartgift.vn>
*/
import { useCallback, useEffect, useRef, useState } from "react";
import { useCallback, useEffect, useLayoutEffect, useRef, useState } from "react";
import { createPortal } from "react-dom";
import { useTranslation } from "react-i18next";
import {
ChevronRight,
@@ -1629,12 +1630,41 @@ function DetailDialog({
);
}
/** "N installs" on a card; a click lists who installed it, when, and where. */
/** "N installs" on a card; a click lists who installed it, when, and where.
* The list floats over the page (a portal at the button, below it or above
* when there is no room, kept inside the window): inside the card list's
* `overflow-hidden` it was cut off at the last card. */
function InstallsButton({ item }: { item: Artifact }) {
const { t } = useTranslation("artifacts");
const [open, setOpen] = useState(false);
const [data, setData] = useState<ArtifactInstalls | null>(null);
const box = useRef<HTMLSpanElement>(null);
const pop = useRef<HTMLDivElement>(null);
const [at, setAt] = useState<{ left: number; top: number } | null>(null);
useLayoutEffect(() => {
if (!open) return setAt(null);
const place = () => {
const b = box.current?.getBoundingClientRect();
const p = pop.current;
if (!b || !p) return;
const w = p.offsetWidth;
const h = p.offsetHeight;
const gap = 4;
const below = b.bottom + gap + h <= window.innerHeight - 8 || b.top - gap - h < 8;
setAt({
left: Math.max(8, Math.min(b.left, window.innerWidth - w - 8)),
top: below ? b.bottom + gap : b.top - gap - h,
});
};
place();
window.addEventListener("resize", place);
window.addEventListener("scroll", place, true);
return () => {
window.removeEventListener("resize", place);
window.removeEventListener("scroll", place, true);
};
}, [open, data]);
useEffect(() => {
if (!open) return;
@@ -1647,7 +1677,8 @@ function InstallsButton({ item }: { item: Artifact }) {
toast.error(err);
});
const onDown = (e: MouseEvent) => {
if (!box.current?.contains(e.target as Node)) setOpen(false);
const n = e.target as Node;
if (!box.current?.contains(n) && !pop.current?.contains(n)) setOpen(false);
};
const onKey = (e: KeyboardEvent) => {
if (e.key === "Escape") setOpen(false);
@@ -1670,66 +1701,70 @@ function InstallsButton({ item }: { item: Artifact }) {
>
{t("installs", { count: item.installs })}
</button>
{open && (
<div
role="dialog"
aria-label={t("installsTitle")}
data-testid="artifact-installs"
className="absolute left-0 z-30 mt-1 w-80 max-w-[80vw] rounded-lg border border-border bg-surface-1 p-2 shadow-xl"
>
<p className="mb-1.5 px-1 text-xs font-medium text-fg-primary">{t("installsTitle")}</p>
{data === null ? (
<p className="px-1 text-xs text-fg-muted">{t("loading")}</p>
) : (
<ul className="max-h-64 space-y-1 overflow-auto">
{data.items.map((i, n) => (
<li key={n} className="rounded px-1.5 py-1 text-xs hover:bg-surface-2">
<div className="flex items-center gap-2">
<span className="min-w-0 flex-1 truncate text-fg-primary">
{i.mine ? t("installsYou") : (i.user ?? t("installsUnknown"))}
{i.replaced && (
<span className="ml-1 text-fg-muted">· {t("installsReplaced")}</span>
)}
</span>
<span className="shrink-0 text-[11px] text-fg-muted">
{new Date(i.at).toLocaleString()}
</span>
</div>
{i.paths && (
<div className="mt-0.5 space-y-0.5">
{i.paths.length ? (
i.paths.map((p) => (
<p
key={p}
className="truncate font-mono text-[10px] text-fg-muted"
title={p}
>
{p}
</p>
))
) : (
<p className="font-mono text-[10px] text-fg-muted">
{t("installsPluginWhere")}
</p>
)}
{open &&
createPortal(
<div
ref={pop}
role="dialog"
aria-label={t("installsTitle")}
data-testid="artifact-installs"
style={at ? { left: at.left, top: at.top } : { left: 0, top: 0, visibility: "hidden" }}
className="fixed z-[60] w-80 max-w-[80vw] rounded-lg border border-border bg-surface-1 p-2 shadow-xl"
>
<p className="mb-1.5 px-1 text-xs font-medium text-fg-primary">{t("installsTitle")}</p>
{data === null ? (
<p className="px-1 text-xs text-fg-muted">{t("loading")}</p>
) : (
<ul className="max-h-64 space-y-1 overflow-auto">
{data.items.map((i, n) => (
<li key={n} className="rounded px-1.5 py-1 text-xs hover:bg-surface-2">
<div className="flex items-center gap-2">
<span className="min-w-0 flex-1 truncate text-fg-primary">
{i.mine ? t("installsYou") : (i.user ?? t("installsUnknown"))}
{i.replaced && (
<span className="ml-1 text-fg-muted">· {t("installsReplaced")}</span>
)}
</span>
<span className="shrink-0 text-[11px] text-fg-muted">
{new Date(i.at).toLocaleString()}
</span>
</div>
)}
</li>
))}
{data.unrecorded > 0 && (
<li className="px-1.5 py-1 text-[11px] text-fg-muted">
{t("installsUnrecorded", { count: data.unrecorded })}
</li>
)}
{data.items.length === 0 && data.unrecorded === 0 && (
<li className="px-1.5 py-1 text-[11px] text-fg-muted">
{t("installsNoneOfYours")}
</li>
)}
</ul>
)}
</div>
)}
{i.paths && (
<div className="mt-0.5 space-y-0.5">
{i.paths.length ? (
i.paths.map((p) => (
<p
key={p}
className="truncate font-mono text-[10px] text-fg-muted"
title={p}
>
{p}
</p>
))
) : (
<p className="font-mono text-[10px] text-fg-muted">
{t("installsPluginWhere")}
</p>
)}
</div>
)}
</li>
))}
{data.unrecorded > 0 && (
<li className="px-1.5 py-1 text-[11px] text-fg-muted">
{t("installsUnrecorded", { count: data.unrecorded })}
</li>
)}
{data.items.length === 0 && data.unrecorded === 0 && (
<li className="px-1.5 py-1 text-[11px] text-fg-muted">
{t("installsNoneOfYours")}
</li>
)}
</ul>
)}
</div>,
document.body
)}
</span>
);
}
@@ -362,6 +362,11 @@ describe("Artifacts", () => {
expect(await within(box).findByText("/h/.claude/skills/deploy")).toBeInTheDocument();
expect(within(box).getByText(/khoa/)).toBeInTheDocument();
expect(within(box).getByText(/1 earlier installs/)).toBeInTheDocument();
// Over the page, outside the card list that clips its overflow — and a
// click inside it keeps it open.
expect(box.parentElement).toBe(document.body);
await user.click(within(box).getByText(/khoa/));
expect(screen.getByTestId("artifact-installs")).toBeInTheDocument();
await user.keyboard("{Escape}");
expect(screen.queryByTestId("artifact-installs")).toBeNull();
});
@@ -7316,7 +7316,7 @@ exports[`screen snapshots > Settings — about 1`] = `
<span
class="text-status-warning/90"
>
UI build v2.8.36 (rebuild client to match)
UI build v2.8.40 (rebuild client to match)
</span>
</p>
</div>
+4 -4
View File
@@ -89,9 +89,9 @@ A dashboard shared by several people runs in **server mode** ([SERVER-MODE.md](S
| `POST /api/auth/login` `{username, password}` | `{user}` + the cookie; `401 EBADLOGIN`, `429 ETOOMANY`, `404 ENOTSERVER` in local mode |
| `POST /api/auth/logout` | `{ok: true}`, cookie cleared |
In server mode a **developer** sees only rows of the workspaces they own (`lanes.owner_id`): every list, count, chart and cost answer is limited to them, and a request naming another user's session, agent, workspace or run (in the path, query or body) answers `404 ENOTFOUND`. Admin and analyst see everything; `GET /api/users` lists every account (`displayName`, `email`, `createdAt`, `logins`, `lastLoginAt` — the last login ever, kept after it ends —, `mustChangePassword` included) with its workspaces, sessions and `today` (since the viewer's local midnight — `?tz_offset=` is the browser's `getTimezoneOffset()`, UTC without it) / 7 / 30-day `{cost, tokens}`, `GET /api/users/storage[?refresh=1]` each user's home size (`{storage: {<id>: {bytes, files, partial, measuredAt} | null}, measuring}` — measured in the background, see [SERVER-MODE.md](SERVER-MODE.md)), `GET /api/users/tmux` each member's own tmux server under one Linux user per account (`{available, installed, matched, users: [{id, username, running, stuck?, version, pid, startedAt, sessions, busy, outdated}]}`; an admin's `POST /api/users/:id/tmux-restart` and `POST /api/users/tmux-restart-idle` restart them with `currentPassword`, audited, `409 ETMUXSTUCK` for a server that will not stop), `GET /api/host` the machine and each user's live share of it (`{at, hostname, os, uptime, cpu: {model, cores, load}, memory: {total, available, used, swapTotal, swapUsed}, disks: [{device, mount, type, kind: "HDD"|"SSD"|null, total, used, free}], homesMount, gpus: [{name, utilization, memoryUsed, memoryTotal}], usage: {cpu, sampleMs, groups: [{key, kind: "user"|"host"|"system", userId?, username?, cpu, memory, memoryPercent, processes, top: [{pid, name, cpu, memory}]}]} | null, bootTime, network: [{name, rx, tx}], diskIo: [{name, read, write}], temperatures: [{chip, label, celsius}], pressure: {cpu, memory, io: {some10, some60, full10} | null} | null (Linux PSI, `/proc/pressure/*`: % of the last 10 / 60 s tasks stalled on it), topProcesses: [{pid, name, cpu, memory, group}], process: {pid, node, uptime, rss, heapUsed}}`, `cpu` also carrying `physicalCores` and `perCore` (busy % per thread), rates in bytes/s over the sample — read-only, cached 3 s, `usage` null and the lists empty off Linux), `GET /api/host/history?range=1h|6h|24h` the CPU/memory history every server on Linux keeps in memory since it started (both read by the Dashboard's System Health tab and open to every signed-in user — `/host` is not staff-only) (`{range, intervalMs, running, startedAt, points: [{at, cpu, memory, memoryUsed, swapUsed, load1, users: {<id>: {cpu, memory}}}]}`, one point per 30 s, averaged to ≤ 360), `GET /api/users/:id/details[?tz_offset=]` one account in full for the Administration page's user detail (`{user (with logins, lastLoginAt), sessions: {total, active, firstAt, lastAt}, usage: {today, last7d, last30d, daily: [{date, cost, tokens}] (30 local days, oldest first), byTool, byModel: [{key, cost, tokens, sessions}]}, workspaces: [{id, title, cwd, tool, kind, branch, status, createdAt, sessions, lastActiveAt, livePanes}], panes: [{id, laneId, laneTitle, laneCwd, tool, model, account, proxy, sessionId, startedAt, status}], recentSessions (15, with cost/tokens), pool: {claude|codex|cursor: {pin, group, accounts: [{name, kind, email, expired, allowed, pinned, quota, quotaPercent, full, livePanes, spend5h, spend7d, userSpend5h, userSpend7d, userTokens5h, userTokens7d, userRuns7d, userLivePanes, userLastAt}]}}, audit}` — `user*` fields are only the runs on that user's workspaces; never a token; `404 ENOUSER`), and `GET /api/users/audit[?limit=]` the staff reads of developers' sessions and admins' account changes (kept `AGENTHUB_AUDIT_DAYS` days, default 180, `0` = forever) (`reader`/`owner` are usernames, `reader_display`/`owner_display` their display names or null) (all `403 EADMIN` for a developer). `GET /api/users` also carries each account's [spend quota](SERVER-MODE.md#spend-quotas) — `quotaSetting` (`{mode: "default"|"unlimited"|"custom", limit_usd?, period?, tz_offset?}`) and `quota` (the one binding them now with `spend`, `pct`, `exceeded`, `resetsAt`, or null) — and the top-level `quotaDefault`; an admin sets the default with `PUT /api/users/quota-default` `{quota: {limit_usd, period, tz_offset} | null, currentPassword}` and a user's own with `PATCH /api/users/:id` `{quota}` (`400 EBADQUOTA`). A workspace owner over their quota gets `429 EQUOTA` from `POST /api/run` and `POST /api/lanes/:id/start`. See [SERVER-MODE.md](SERVER-MODE.md#what-each-user-sees).
In server mode a **developer** sees only rows of the workspaces they own (`lanes.owner_id`): every list, count, chart and cost answer is limited to them, and a request naming another user's session, agent, workspace or run (in the path, query or body) answers `404 ENOTFOUND`. Admin and analyst see everything; `GET /api/users` lists every account (`displayName`, `email`, `createdAt`, `logins`, `lastLoginAt` — the last login ever, kept after it ends —, `mustChangePassword` included) with its workspaces, sessions and `today` (since the viewer's local midnight — `?tz_offset=` is the browser's `getTimezoneOffset()`, UTC without it) / 7 / 30-day `{cost, tokens}`, `GET /api/users/storage[?refresh=1]` each user's home size (`{storage: {<id>: {bytes, files, partial, measuredAt} | null}, measuring}` — measured in the background, see [SERVER-MODE.md](SERVER-MODE.md)), `GET /api/users/tmux` each member's own tmux server under one Linux user per account (`{available, installed, matched, users: [{id, username, running, stuck?, version, pid, startedAt, sessions, busy, outdated}]}`; an admin's `POST /api/users/:id/tmux-restart` and `POST /api/users/tmux-restart-idle` restart them with `currentPassword`, audited, `409 ETMUXSTUCK` for a server that will not stop), `GET /api/host` the machine and each user's live share of it (`{at, hostname, os, uptime, cpu: {model, cores, load}, memory: {total, available, used, swapTotal, swapUsed}, disks: [{device, mount, type, kind: "HDD"|"SSD"|null, total, used, free}], homesMount, gpus: [{name, utilization, memoryUsed, memoryTotal}], usage: {cpu, sampleMs, groups: [{key, kind: "user"|"host"|"system", userId?, username?, displayName?, avatarAt?, cpu, memory, memoryPercent, processes, top: [{pid, name, cpu, memory}]}]} | null, bootTime, network: [{name, rx, tx}], diskIo: [{name, read, write}], temperatures: [{chip, label, celsius}], pressure: {cpu, memory, io: {some10, some60, full10} | null} | null (Linux PSI, `/proc/pressure/*`: % of the last 10 / 60 s tasks stalled on it), topProcesses: [{pid, name, cpu, memory, group}], process: {pid, node, uptime, rss, heapUsed}}`, `cpu` also carrying `physicalCores` and `perCore` (busy % per thread), rates in bytes/s over the sample — read-only, cached 3 s, `usage` null and the lists empty off Linux), `GET /api/host/history?range=1h|6h|24h` the CPU/memory history every server on Linux keeps in memory since it started (both read by the Dashboard's System Health tab and open to every signed-in user — `/host` is not staff-only) (`{range, intervalMs, running, startedAt, points: [{at, cpu, memory, memoryUsed, swapUsed, load1, users: {<id>: {cpu, memory}}}]}`, one point per 30 s, averaged to ≤ 360), `GET /api/users/:id/details[?tz_offset=]` one account in full for the Administration page's user detail (`{user (with logins, lastLoginAt), sessions: {total, active, firstAt, lastAt}, usage: {today, last7d, last30d, daily: [{date, cost, tokens}] (30 local days, oldest first), byTool, byModel: [{key, cost, tokens, sessions}]}, workspaces: [{id, title, cwd, tool, kind, branch, status, createdAt, sessions, lastActiveAt, livePanes}], panes: [{id, laneId, laneTitle, laneCwd, tool, model, account, proxy, sessionId, startedAt, status}], recentSessions (15, with cost/tokens), pool: {claude|codex|cursor: {pin, group, accounts: [{name, kind, email, expired, allowed, pinned, quota, quotaPercent, full, livePanes, spend5h, spend7d, userSpend5h, userSpend7d, userTokens5h, userTokens7d, userRuns7d, userLivePanes, userLastAt}]}}, audit}` — `user*` fields are only the runs on that user's workspaces; never a token; `404 ENOUSER`), and `GET /api/users/audit[?limit=]` the staff reads of developers' sessions and admins' account changes (kept `AGENTHUB_AUDIT_DAYS` days, default 180, `0` = forever) (`reader`/`owner` are usernames, `reader_display`/`owner_display` their display names or null) (all `403 EADMIN` for a developer). `GET /api/users` also carries each account's [spend quota](SERVER-MODE.md#spend-quotas) — `quotaSetting` (`{mode: "default"|"unlimited"|"custom", limit_usd?, period?, tz_offset?}`) and `quota` (the one binding them now with `spend`, `pct`, `exceeded`, `resetsAt`, or null) — and the top-level `quotaDefault`; an admin sets the default with `PUT /api/users/quota-default` `{quota: {limit_usd, period, tz_offset} | null, currentPassword}` and a user's own with `PATCH /api/users/:id` `{quota}` (`400 EBADQUOTA`). A workspace owner over their quota gets `429 EQUOTA` from `POST /api/run` and `POST /api/lanes/:id/start`. See [SERVER-MODE.md](SERVER-MODE.md#what-each-user-sees).
An admin changes accounts with `POST /api/users` (`{username, password, role}` → `201 {user}`; the password may be any non-empty starting one and the account gets `mustChangePassword: true`), `GET /api/users/:id` (`{user, workspaces: [{id, title, cwd}]}`, the account with `logins` and `lastLoginAt`, for the Edit form; admin or analyst), `GET /api/users/:id/avatar` (that account's picture for the sidebar's owner groups — admin or analyst, `nosniff` + `sandbox` CSP, `404 ENOAVATAR`; a workspace in `GET /api/lanes` carries `owner_avatar_at` to version its URL), `PATCH /api/users/:id` (any of `{displayName?, email?, role?, disabled?, password?, addWorkspaces?: [laneId], removeWorkspaces?: [laneId], signOut?: true}` — every field validated before any is written; `removeWorkspaces` leaves the workspace with no owner and answers `409 ENOTOWNED` for one the user does not own; an admin's own `password`/`signOut` keeps the login it came from; a password for SOMEONE ELSE may be anything non-empty and sets their `mustChangePassword`, the admin's OWN must pass every rule below, `400 EWEAKPASSWORD` with `rules`), `POST /api/users/:id/password` (`{password}`; that user's logins end; the same two bars) `POST /api/users/:id/workspaces` (`{laneId}`; the workspace's owner becomes that user) and `DELETE /api/users/:id` (delete for good: the user's logins, panes, workspaces' records, prefs and — under S3 — their Linux account and whole home, removed FIRST so a refusal (`502 EPROVISION`) leaves every row; `409 ESELF` for yourself, `409 ELASTADMIN` for the last admin; session history and the audit log stay). No path — this API or the `agenthub users` CLI — may demote, disable or delete the last enabled admin (`409 ELASTADMIN`). Every one carries `currentPassword`, the admin's own password: `403 EREAUTH` when it is wrong or missing, `429 ETOOMANY` after 10 wrong tries in 15 minutes. `409 ESELF` for an admin's own role or disable, `409 EEXISTS` for a taken name, `400 EBADUSER`/`EBADPASSWORD`/`EBADROLE`/`EBADEMAIL`/`EBADFIELD`, `404 ENOUSER`/`ENOLANE`; a developer or analyst gets `403 EADMIN`, and local mode `403 ELOCALMODE`. No answer carries a password hash; each change is written to the audit log.
An admin changes accounts with `POST /api/users` (`{username, password, role}` → `201 {user}`; the password may be any non-empty starting one and the account gets `mustChangePassword: true`), `GET /api/users/:id` (`{user, workspaces: [{id, title, cwd}]}`, the account with `logins` and `lastLoginAt`, for the Edit form; admin or analyst), `GET /api/users/:id/avatar` (that account's picture for the sidebar's owner groups — admin or analyst, `nosniff` + `sandbox` CSP, `404 ENOAVATAR`; a workspace in `GET /api/lanes` carries `owner_avatar_at` to version its URL), `PATCH /api/users/:id` (any of `{displayName?, email?, role?, disabled?, password?, addWorkspaces?: [laneId], removeWorkspaces?: [laneId], signOut?: true}` — every field validated before any is written; `removeWorkspaces` leaves the workspace with no owner and answers `409 ENOTOWNED` for one the user does not own; an admin's own `password`/`signOut` keeps the login it came from; a password for SOMEONE ELSE may be anything non-empty and sets their `mustChangePassword`, the admin's OWN must pass every rule below, `400 EWEAKPASSWORD` with `rules`), `POST /api/users/:id/password` (`{password}`; that user's logins end; the same two bars) `POST /api/users/:id/workspaces` (`{laneId}`; the workspace's owner becomes that user) and `DELETE /api/users/:id` (delete for good: the user's logins, panes, workspaces' records, prefs and — under S3 — their Linux account and whole home, removed FIRST so a refusal (`502 EPROVISION`) leaves every row; `409 ESELF` for yourself, `409 ELASTADMIN` for the last admin; session history and the audit log stay). No path — this API or the `agenthub users` CLI — may demote, disable or delete the last enabled admin (`409 ELASTADMIN`). Every one carries `currentPassword`, the admin's own password: `403 EREAUTH` when it is wrong or missing, `429 ETOOMANY` after 10 wrong tries in 15 minutes. `409 ESELF` for an admin's own role or disable, `409 EEXISTS` for a taken name, `409 ENAMETAKEN` for a `displayName` that reads as another account's (the same rule as `PATCH /api/me/account`), `400 EBADUSER`/`EBADPASSWORD`/`EBADROLE`/`EBADEMAIL`/`EBADFIELD`, `404 ENOUSER`/`ENOLANE`; a developer or analyst gets `403 EADMIN`, and local mode `403 ELOCALMODE`. No answer carries a password hash; each change is written to the audit log.
`GET /api/accounts[?force=1]` lists the Claude account pool ([SERVER-MODE.md](SERVER-MODE.md#claude-account-pool)): `{accounts: [{name, addedAt, expiresAt, expired, livePanes, spend5h, spend7d, tokens5h, tokens7d, quota: {fiveHour, sevenDay, checkedAt, error} | null, quotaPercent, full}], codex: [{name, kind: "apikey"|"chatgpt", addedAt, livePanes, spend5h, spend7d, tokens5h, tokens7d, quota: {windows: [{label, usedPercent, resetsAt}], asOf} | null, quotaPercent, full}], pins: {claude: [{userId, username, account, pinnedAt}], codex: […]}, blockPercent}` — `pins` are the users an admin pinned to one account (empty in local mode, keyed by tool: `claude`, `codex`, `cursor`); `codex` is the [Codex account pool](SERVER-MODE.md#codex-account-pool); `cursor: [{name, kind: "apikey", addedAt, livePanes, …}]` is the [Cursor account pool](SERVER-MODE.md#cursor-account-pool) (API keys; `quota` null, spend 0); `groups: [{id, name, createdAt, users: [{userId, username}], accounts: {claude: [names], codex: […], cursor: […]}}]` are the [pool groups](SERVER-MODE.md#account-pool-groups) (empty in local mode). `spend*`/`tokens*` = the sessions the account's runs held, over the hours they held them (a Codex session counts whole once its run started in the window). It never includes a key or a token. `force=1` re-reads every quota first. Run rows (`/api/run/history`) carry `account`, the pool account their pane ran on (`null` = the machine's own login).
@@ -110,7 +110,7 @@ An admin (server mode) adds an account with `POST /api/accounts` `{tool: "claude
| `PUT /api/me/pool-accounts` `{tool, account: "<name>" \| null}` | the account's new sessions of `tool` run on `account` (`null` = automatic): `{tool, choice, switched}` — `switched` as for an admin's pin (a Claude login goes into the home at once). Below an admin's pin (`409 ELOCKED`), only an account the group allows (`404 ENOACCOUNT`), `400 EBADTOOL`; login cookie only, audited |
| `POST /api/me/onboarded` | the first-run setup is done or skipped (kept from the first call) |
| `GET /api/me/account` | Settings → Account. Server mode: `{mode: "server", user: {id, username, role, disabled, displayName, email, avatarAt, createdAt}, workspaces, sessions, logins, lastLoginAt}`; local mode `{mode: "local"}` |
| `PATCH /api/me/account` `{displayName?, email?}` | the account's own profile (`""`/`null` clears), the same shape back; `400 EBADFIELD` (display name over 100 characters or with control characters) / `EBADEMAIL` |
| `PATCH /api/me/account` `{displayName?, email?}` | the account's own profile (`""`/`null` clears), the same shape back; `400 EBADFIELD` (display name over 100 characters or with control characters) / `EBADEMAIL`; `409 ENAMETAKEN` for a display name that reads as another account's display name or username (marks, case, punctuation, bracketed parts like "(you)" and look-alike letters ignored; keeping the name you have is always allowed) |
| `POST /api/me/password` `{currentPassword, newPassword}` | `{ok: true}`; every other login of the account ends, this one stays, and `mustChangePassword` clears. The new password must pass every rule — at least 8 characters, a lower-case and an upper-case letter, a digit, a symbol, not containing the username — and differ from the current one: `400 EWEAKPASSWORD` with `error.rules` (`length`, `lower`, `upper`, `digit`, `symbol`, `username`, `reused`). `403 EREAUTH` wrong current password (`429 ETOOMANY` after 10), `400 EBADPASSWORD` |
| `POST /api/me/sign-out-others` | `{ended}` — every login of the account but this one ends |
| `GET /api/me/avatar` | server mode: the account's own picture (`nosniff`, `sandbox` CSP; the client adds `?v=<avatarAt>`), `404 ENOAVATAR` when it has none |
@@ -3036,7 +3036,7 @@ POST /api/tool-profiles/:tool/activate Body: { name }
DELETE /api/tool-profiles/:tool Body: { name }
```
The login profiles of the AI tools other than Claude Code (`server/lib/tool-profiles.js`; Claude's are `/api/cc-config/profiles` above). Settings → AI tools → **Profiles** shows both, one tab per tool. `:tool` is `codex` (the whole `$CODEX_HOME/auth.json`: a ChatGPT login or an API key) or `omp` (every credential Oh My Pi stores, read with `omp auth export` and written back with `omp auth import -` — secrets on stdin, never argv; import merges, so a switch first empties omp's credential tables in `~/.omp/agent/agent.db`, and a failed import puts the outgoing set back). `antigravity` is listed with `supported: false` and a `reason`: agy keeps its login in the OS keyring, which AgentHub does not touch (`409 EUNSUPPORTED`). `GET /` answers `{tools: [{tool, supported, reason?, items, active, error?}]}`, each item `{name, label, savedAt, active, kind, email, detail}` — never a token; one tool's failure (omp not installed) is that tool's `error`, not the request's. Profiles are stored one file each under `<CLAUDE_HOME>/agenthub-profiles/<tool>/` (0700/0600) with cc-profiles' rules: one login is one profile (`409 EDUPLICATE`), a switch re-saves the outgoing login into its own profile first and refuses one never saved (`409 EUNSAVED`), the live login's profile is not deletable (`409 EACTIVE`), `409 ENOAUTH` when the tool holds no login, `409 ENOTINSTALLED`, `502 EOMP` when omp's own command fails. Writes are same-origin only; in server mode `/tool-profiles` is in `ADMIN_WRITE_PREFIXES` and `STAFF_ONLY_PREFIXES` — it changes the HOST's own logins.
The login profiles of the AI tools other than Claude Code (`server/lib/tool-profiles.js`; Claude's are `/api/cc-config/profiles` above). Settings → AI tools → **Profiles** shows both, one tab per tool. `:tool` is `codex` (the whole `$CODEX_HOME/auth.json`: a ChatGPT login or an API key) or `omp` (every credential Oh My Pi stores, read with `omp auth export` and written back with `omp auth import -` — secrets on stdin, never argv; import merges, so a switch first empties omp's credential tables in `~/.omp/agent/agent.db`, and a failed import puts the outgoing set back). `antigravity` is listed with `supported: false` and a `reason`: agy keeps its login in the OS keyring, which AgentHub does not touch (`409 EUNSUPPORTED`). `GET /` answers `{tools: [{tool, supported, reason?, items, active, error?}]}`, each item `{name, label, savedAt, active, kind, email, detail}` — never a token; one tool's failure (omp not installed) is that tool's `error`, not the request's. Profiles are stored one file each under `<CLAUDE_HOME>/agenthub-profiles/<tool>/` (0700/0600) with cc-profiles' rules: one login is one profile (`409 EDUPLICATE`), a switch re-saves the outgoing login into its own profile first and refuses one never saved (`409 EUNSAVED`), and for Codex then runs `codex app-server daemon restart` when that CODEX_HOME's shared daemon is running (its `app-server-control.sock` resolves) — the daemon holds the login it started with, so without it neither an open nor a NEW Codex session saw the switch; open TUIs reconnect on their own — answering `following: true` (the activate response's extra field; `false` when no daemon ran), the live login's profile is not deletable (`409 EACTIVE`), `409 ENOAUTH` when the tool holds no login, `409 ENOTINSTALLED`, `502 EOMP` when omp's own command fails. Writes are same-origin only; in server mode `/tool-profiles` is in `ADMIN_WRITE_PREFIXES` and `STAFF_ONLY_PREFIXES` — it changes the HOST's own logins.
### AI Proxy profiles
+1 -1
View File
File diff suppressed because one or more lines are too long
+7 -1
View File
@@ -372,7 +372,13 @@ Wherever the dashboard names a person — the sidebar's sign-out tooltip, the
workspace list's owner groups, the Users page and its audit log, the Edit user
form — it shows their display name when one is set, else the username (the
Users page keeps `@username` beside a display name). Logging in, the CLI and
the Linux account always use the username.
the Linux account always use the username. No one can pose as a colleague: a
display name that reads as another account's display name or username — marks,
case, spaces, punctuation, a bracketed part like "(you)" and look-alike letters
from other alphabets ignored — is refused (`409 ENAMETAKEN`), whether the user
sets it in Settings → Account or an admin does on the Users page. A name an
account already has is kept on an unrelated save; rename an old clash from the
Users page.
Under S3 an admin's folder browser opens on their own home too; ticking
**Server folder** opens the host account's home (`/home/<host>`), where a
+2 -2
View File
@@ -1,12 +1,12 @@
{
"name": "agenthub",
"version": "2.8.36",
"version": "2.8.40",
"lockfileVersion": 3,
"requires": true,
"packages": {
"": {
"name": "agenthub",
"version": "2.8.36",
"version": "2.8.40",
"hasInstallScript": true,
"license": "UNLICENSED",
"dependencies": {
+1 -1
View File
@@ -1,6 +1,6 @@
{
"name": "agenthub",
"version": "2.8.36",
"version": "2.8.40",
"bin": {
"agenthub": "bin/agenthub.js"
},
@@ -222,6 +222,26 @@ describe("Settings → Account (/api/me)", () => {
assert.equal(bad.body.error.code, "EBADEMAIL");
});
it("refuses a display name that reads as another account's (ENAMETAKEN)", async () => {
assert.equal(auth.nameKey("Nguyễn Ngọc Trí Vĩ (you)"), auth.nameKey("nguyen.ngoc-tri VI"));
assert.equal(auth.nameKey("Đào"), "dao");
// Cyrillic о and а in place of Latin letters.
assert.equal(auth.nameKey("Mіа Trоn"), auth.nameKey("Mia Tron"));
const leo = await login("leo", "member-pass-2");
// mia's display name is "Mia Tran"; boss has none, only a username.
for (const name of ["MIA TRẦN (you)", "Mia\u200b Tran", "mia", "Bóss"]) {
const r = await call("PATCH", "/api/me/account", { cookie: leo, body: { displayName: name } });
assert.equal(r.status, 409, name);
assert.equal(r.body.error.code, "ENAMETAKEN");
}
assert.equal(auth.findUser("leo").display_name, "Leo N");
// Keeping one's own name, or a name of one's own, still works.
for (const name of ["Leo N", "Leo Nguyen"]) {
const r = await call("PATCH", "/api/me/account", { cookie: leo, body: { displayName: name } });
assert.equal(r.status, 200, name);
}
});
it("changes its own password only with the old one, keeping this login", async () => {
const leo = await login("leo", "member-pass-2");
const other = await login("leo", "member-pass-2");
+27
View File
@@ -97,6 +97,33 @@ describe("tool-profiles: codex", () => {
);
});
it("restarts a running Codex daemon so open sessions follow; none running, nothing to restart", async () => {
login("acct-a", "alice@example.com");
profiles.saveProfile("codex");
login("acct-b", "bob@example.com");
profiles.saveProfile("codex");
assert.equal(profiles.activateProfile("codex", "alice").following, false);
const log = path.join(ROOT, "codex-argv.log");
const bin = path.join(ROOT, "fake-codex.sh");
fs.writeFileSync(bin, `#!/bin/sh\necho "$CODEX_HOME $*" >> '${log}'\n`, { mode: 0o755 });
const sock = path.join(process.env.CODEX_HOME, "app-server-control", "app-server-control.sock");
fs.mkdirSync(path.dirname(sock), { recursive: true });
fs.writeFileSync(sock, "");
process.env.AGENTHUB_CODEX_BIN = bin;
try {
assert.equal(profiles.activateProfile("codex", "bob").following, true);
for (let i = 0; i < 50 && !fs.existsSync(log); i++)
await new Promise((r) => setTimeout(r, 20));
assert.equal(
fs.readFileSync(log, "utf8").trim(),
`${process.env.CODEX_HOME} app-server daemon restart`
);
} finally {
delete process.env.AGENTHUB_CODEX_BIN;
}
});
it("refuses to switch away from an unsaved login", () => {
login("acct-a", "alice@example.com");
profiles.saveProfile("codex");
+6
View File
@@ -93,6 +93,12 @@ describe("GET /api/sessions/:id/stats tool totals", () => {
assert.equal(body.tool_call_count, 2);
assert.equal(body.distinct_tool_count, 1);
});
it("counts each call once in tools_used, not its Pre and Post events", async () => {
const one = (await get("/api/sessions/ua-1/stats")).body.tools_used;
assert.ok(one.every((t) => t.count === 1));
const two = (await get("/api/sessions/ua-2/stats")).body.tools_used;
assert.deepEqual(two, [{ tool_name: "Read", count: 2 }]);
});
});
describe("GET /api/events labels", () => {
+6
View File
@@ -2547,10 +2547,16 @@ const stmts = {
FROM events
WHERE session_id = ? AND tool_name IS NOT NULL
`),
// One row per call, like sessionToolCallTotals: a call is its PreToolUse,
// or its PostToolUse in a session that has none (imported). Counting every
// event with a tool_name counted each call twice.
sessionToolUsageCounts: db.prepare(`
SELECT tool_name, COUNT(*) as count
FROM events
WHERE session_id = ? AND tool_name IS NOT NULL
AND event_type = CASE
WHEN EXISTS (SELECT 1 FROM events WHERE session_id = ? AND event_type = 'PreToolUse')
THEN 'PreToolUse' ELSE 'PostToolUse' END
GROUP BY tool_name
ORDER BY count DESC
LIMIT 15
+45 -1
View File
@@ -20,7 +20,8 @@
* admin on the Users page (`server/routes/users.js`), where every change
* needs the admin's own password again (`confirmPassword`, throttled).
* An account also carries an optional display name and email
* (`validateProfile`), which its owner edits in Settings → Account without
* (`validateProfile`; no two accounts may read as one name — `assertNameFree`,
* ENAMETAKEN), which its owner edits in Settings → Account without
* that step-up; changing their OWN password there needs the old one and
* keeps the login it was changed from (`setPassword`'s `keepToken`).
* - Two password bars: an admin may GIVE any non-empty password (a create or a
@@ -247,6 +248,47 @@ function validateProfile(body) {
return out;
}
// Latin look-alikes from other scripts, so "Vĩ" in Cyrillic still collides.
const CONFUSABLE = {
а: "a", в: "b", е: "e", к: "k", м: "m", н: "h", о: "o", р: "p", с: "c", т: "t",
у: "y", х: "x", і: "i", ј: "j", ԁ: "d", ѕ: "s", ӏ: "l", α: "a", β: "b", ε: "e",
η: "n", ι: "i", κ: "k", ν: "v", ο: "o", ρ: "p", τ: "t", υ: "u", χ: "x", "ı": "i",
};
/**
* The name two accounts may not share: no marks (Vĩ = Vi), đ = d, any case,
* look-alike letters folded, and only letters and digits outside brackets —
* so "Nguyễn Ngọc Trí Vĩ (you)" and "nguyen.ngoc-tri vi" are both one name.
*/
function nameKey(s) {
return String(s || "")
.normalize("NFKD")
.replace(/\p{M}/gu, "")
.toLowerCase()
.replace(/\([^)]*\)|\[[^\]]*\]/g, "")
.replace(/[đð]/g, "d")
.replace(/./gu, (c) => CONFUSABLE[c] ?? c)
.replace(/[^\p{L}\p{N}]/gu, "");
}
/**
* Refuse (ENAMETAKEN) a display name that reads as another account's display
* name or username, so no one can pose as a colleague. Keeping the name one
* already has is always allowed — a clash from before this rule is no reason
* to block an admin's unrelated edit.
*/
function assertNameFree(userId, fields) {
if (!("display_name" in fields) || !fields.display_name) return;
const key = nameKey(fields.display_name);
if (!key) return;
const own = sql("SELECT display_name FROM users WHERE id = ?").get(userId);
if (own && own.display_name === fields.display_name) return;
const rows = sql("SELECT id, username, display_name FROM users WHERE id != ?").all(userId);
if (rows.some((u) => nameKey(u.username) === key || nameKey(u.display_name) === key)) {
throw err("ENAMETAKEN", "displayName: another account already uses that name");
}
}
/** Write `validateProfile`'s result for one account. */
function setProfile(userId, fields) {
for (const [col, value] of Object.entries(fields)) {
@@ -812,6 +854,8 @@ module.exports = {
unassignWorkspace,
ownedWorkspaces,
validateProfile,
nameKey,
assertNameFree,
setProfile,
validatePassword,
validateOwnPassword,
+2 -1
View File
@@ -555,12 +555,13 @@ async function usage(t0, t1, p0, p1, mem) {
const round = (n) => Math.round(n * 10) / 10;
/** A user's group fields: id, username and display name (every viewer reads them). */
/** A user's group fields: id, username, display name and avatar version (every viewer reads them). */
const userExtra = (u) => ({
kind: "user",
userId: u.id,
username: u.username,
displayName: u.displayName ?? null,
avatarAt: u.avatarAt ?? null,
});
async function collect() {
+35 -2
View File
@@ -6,7 +6,9 @@
* swap it back in later, touching nothing but that login.
*
* - `codex`: the whole `$CODEX_HOME/auth.json` (a ChatGPT login or an API
* key). Config, sessions, skills and history stay where they are.
* key). Config, sessions, skills and history stay where they are. A
* switch then restarts Codex's shared app-server daemon when one runs,
* so running and new sessions follow the new login (`follow`).
* - `omp`: every credential Oh My Pi stores, read with `omp auth export`
* and put back with `omp auth import -` (secrets on stdin, never argv).
* Import MERGES, so a switch first empties omp's own credential tables in
@@ -85,6 +87,15 @@ const codex = {
writeJson(codexAuthFile(), secret);
return [codexAuthFile()];
},
// Codex ≥ 0.164 talks to the API through a shared app-server daemon that
// outlives every TUI and holds the login it started with: a new auth.json
// reached neither a running session nor a NEW one, which kept the old
// (often out-of-quota) account. Restarting the daemon makes it read the new
// login, and every open TUI reconnects to it on its own (verified on
// 0.164.0). No daemon running: nothing to restart — the next one reads the file.
follow() {
return restartCodexDaemon();
},
identify(auth) {
if (auth.tokens && auth.tokens.id_token) {
const c = jwtClaims(auth.tokens.id_token) || {};
@@ -107,6 +118,26 @@ const codex = {
},
};
/** The control socket's link points into a per-user runtime dir that exists only while a daemon runs. */
const codexDaemonSocket = () =>
path.join(getCodexHome(), "app-server-control", "app-server-control.sock");
/** Restart the CODEX_HOME's app-server daemon when one is running; true when a restart was started. */
function restartCodexDaemon() {
if (!fs.existsSync(codexDaemonSocket())) return false;
const bin = process.env.AGENTHUB_CODEX_BIN || require("./ai-tools").detect("codex").path;
if (!bin) return false;
cp.execFile(
bin,
["app-server", "daemon", "restart"],
{ env: { ...process.env, CODEX_HOME: getCodexHome() }, timeout: 60000, windowsHide: true },
(e) => {
if (e) console.warn(`[tool-profiles] codex daemon restart failed: ${e.message}`);
}
);
return true;
}
// ── Oh My Pi ───────────────────────────────────────────────────────────
const ompDir = () => process.env.AGENTHUB_OMP_DIR || path.join(os.homedir(), ".omp");
@@ -405,7 +436,9 @@ function activateProfile(tool, name) {
}
const target = readProfile(tool, name);
const written = a.write(target.secret, live);
return { ok: true, name, savedOutgoing, written };
// `following`: running sessions take the new login without a restart of their own.
const following = typeof a.follow === "function" ? a.follow() : false;
return { ok: true, name, savedOutgoing, written, following };
}
function deleteProfile(tool, name) {
+5
View File
@@ -489,6 +489,11 @@ const paths = {
items: { type: "string" },
description: "Files the switch wrote.",
},
following: {
type: "boolean",
description:
"Codex: a running app-server daemon was restarted, so open and new sessions use the new login.",
},
},
},
"Switched."
+2
View File
@@ -197,6 +197,7 @@ const { recordAccountChange } = require("../lib/audit");
const ACCOUNT_STATUS = {
EBADFIELD: 400,
EBADEMAIL: 400,
ENAMETAKEN: 409,
EBADPASSWORD: 400,
EWEAKPASSWORD: 400,
EBADIMAGE: 400,
@@ -262,6 +263,7 @@ router.patch(
sameOriginGuard,
accountWrite(async (_req, user, body) => {
const fields = auth.validateProfile({ displayName: body.displayName, email: body.email });
auth.assertNameFree(user.id, fields);
auth.setProfile(user.id, fields);
return { out: accountView(user.id), action: "own profile" };
})
+1 -1
View File
@@ -702,7 +702,7 @@ router.get("/:id/file", async (req, res) => {
function sessionStats(sessionId) {
const totalEvents = stmts.sessionEventCount.get(sessionId)?.count ?? 0;
const eventsByType = stmts.sessionEventTypeCounts.all(sessionId);
const tools = stmts.sessionToolUsageCounts.all(sessionId);
const tools = stmts.sessionToolUsageCounts.all(sessionId, sessionId);
// Exact totals: `tools_used` is a top-15 list, so its sum is not the count.
// A tool call is a PreToolUse; a session with none (imported) counts PostToolUse.
const toolTotals = stmts.sessionToolCallTotals.get(sessionId) || {};
+2
View File
@@ -171,6 +171,7 @@ router.get("/tmux", (_req, res) => {
const STATUS = {
EBADUSER: 400,
EBADEMAIL: 400,
ENAMETAKEN: 409,
ENOTOWNED: 409,
EBADPASSWORD: 400,
EWEAKPASSWORD: 400,
@@ -562,6 +563,7 @@ router.patch(
// 1. Validate everything — nothing is written until all of it passes.
const profile = auth.validateProfile(body);
auth.assertNameFree(u.id, profile);
const roleChange = body.role !== undefined && body.role !== u.role;
if (roleChange) {
if (self) throw ESELF("you cannot change your own role");