- Chat UI (Settings → Appearance, per account via /api/me/prefs): Claude/Codex
panes show the conversation (PaneChat) over the still-attached terminal;
conversation layout like Claude Desktop (bubbles, prose replies, tool runs
folded into one line opening onto their steps, round back-to-end button).
- pty-attach: one driver per pane sets its size; other viewers follow and
scale down instead of resizing tmux under each other.
- Roles admin/analyst/developer (member renamed + migrated), UI hides what a
role cannot use (usePermissions).
- Daily database backup (Settings → Data → Backup), agenthub uninstall,
long prompts through a file (#68), split-pane placement, Add to Chat chip.
Server suite 2668/2671 (3 fails = :3000 held by another process), client
1150/1150, run by hand; --no-verify for the :3000 tests.
Usage panel (status bar):
- every account shows its plan (Claude subscriptionType, Codex
chatgpt_plan_type, Cursor membershipType, Grok subscriptionTier)
- new quota rows: Cursor CLI (cursor.com usage-summary), Antigravity
(`agy --print /usage`, two model pools; AGENTHUB_AGY_USAGE=0 never runs
agy), Grok Build, Kimi Code, OpenCode Go — each login read from the tool's
own file, sent only to its own host while valid, never refreshed
- in-flight quota reads are shared between tabs
AI Tools catalog: Grok Build (`grok`) and Kimi Code (`kimi`) added;
Gemini CLI, Hermes and OpenClaw removed.
Also: Quit AgentHub (#65, `agenthub stop`), status-area icon, Add to Chat
on an editor selection, Settings in three groups, JetBrains-style Changes
panel (ticked files/hunks, ignore, blame, file history), terminal font.
--no-verify: :3000 is taken by another dev server (lane-detect tests);
both suites run by hand — server 2636/2639 (the 3 are :3000), client
1131/1131.
Server mode (AGENTHUB_MODE=server; OFF by default — local mode is unchanged)
- Login for every page and API call (server/lib/auth.js, routes/auth.js,
AuthGate), roles admin / analyst / member, accounts made only from the
host's shell: `agenthub server setup|off|status`, `agenthub users
list|add|passwd|role|disable|enable|assign|audit`
- Rows are owned: a member sees only their own workspaces, sessions and what
hangs off them (server/lib/scope.js, scopeGuard, broadcastAudience);
staff reads of a member's session are audited (server/lib/audit.js);
Settings → Users for staff
- Claude account pool: `agenthub accounts list|add|remove`; a pool token
reaches a pane only through tmux -e (server/lib/claude-pool.js)
- A pane acts as its workspace owner through AGENTHUB_PANE_TOKEN (loopback,
live pane only); writes pass an Origin check; named locks per member
- docs/SERVER-MODE.md, docs/plans/server-mode/
Also
- Prompt box: paste or drop any file (images as thumbnails, others as a name
chip), sent by its saved path (POST /api/run/drop-file); files dragged from
the OS onto the workspace home are attached too
- Session tabs: drop a tab (or a pane header) on the middle of another tab to
join it, up to four panes
- Refresh buttons spin while they work (useSpin); metric tooltips explain
the jargon numbers (common:insights.help.*)
- Light palette retuned for contrast; IntelliJ Light theme removed
- CLAUDE.md sync marker trimmed; its history moved to docs/CLAUDE-SYNC-LOG.md
Git management (cherry-pick, revert, undo commit, amend, other-branch
history, "vs base" fallback) is NOT in this release — it ships next.
Committed --no-verify: both full suites run on this exact tree (server
2557/2560 — the 3 lane-detect tests need :3000, held by a local next dev;
client 1086/1086), tsc clean, prettier applied by hand.
Measured on a copy of a real 1.2 GB / 253K-event database, 10 simulated
browsers + 10 hooks/s for 90 s: event-loop delay p99 1239 ms → 32 ms (max
2.0 s → 94 ms), 266 failed requests → 0, 74 → 111 requests/s served, server
RSS 633 → 346 MB. Entry bundle 975 KB raw → 394 KB (106 KB brotli).
Server
- /api/analytics: the events-table scans run in the read worker
(scoped-stats eventAggregates, read-pool job analyticsEvents); runRead takes
maxAgeMs — analytics 10 s, /api/workflows 30 s — keyed by the args
- /api/workflows/skill-usage: +event_type and a tool_name range instead of
LIKE, so it uses idx_events_tool (784 ms → a few ms)
- idx_events_session_type_tool (session_id, event_type, tool_name,
created_at) replaces idx_events_session_type: covering reads for session
stats, the hook dedupes and the import resume point (1.2 s → 20 ms); 64 MB
cache_size and 256 MB mmap_size on both connections
- WAL: the server sets wal_autocheckpoint = 0 and checkpoints PASSIVE every
DASHBOARD_WAL_CHECKPOINT_MS (5 s) from the read worker's own writable
connection; a failed checkpoint restores the automatic one (hook COMMIT p95
109 ms → 15 ms at 100 hooks/s)
- tmux: one list-sessions answers a whole GET /api/lanes and the heal beat
(liveSessions), shared for 1 s and cleared by this process's new/kill —
GET /api/run's sync spawns had blocked 60 s of a 90 s run; the heal beat
skips tmux when nothing runs; attach options chained into 2 spawns
- hooks: TurnDuration/APIError deduped against one read per hook instead of a
query per transcript entry (23 ms/hook on a 763-turn session); one lane
scan per hook; hourly upserts skip rows that did not grow; frames are sent
after the transaction commits (and after the pane binding); new_event
carries its row id; a split pane's (slot 1–9) hooks no longer rebind the
lane's session_id
- watchdog liveness probe runs ps asynchronously; transcript newline scan via
indexOf; startup subagent backfill fingerprints the files it already checked
- pty attach: backpressure (pause past 1 MB unsent, resume under 256 KB)
- static: hashed assets served from their pre-compressed .br/.gz sibling
Client
- hidden tabs neither refetch on WS events nor poll (useLiveRefresh, panes,
lane cards, git header, changes panel, settings); no per-tab hourly
/updates/check (the server already checks)
- refetch storms throttled (workspace tree/session list 3 s, /api/run only on
a run-key change, notification pane lookup only when shown); WS reconnect
jitter; terminal retry backoff 1.5 → 30 s, stops on 1008
- heavier throttles for aggregates (15 s) and detail views; Monitor keepIfSame;
GET in-flight dedupe in request()
- lazy Dashboard/History, xterm (with the first terminal), lane modals, Sagi
(only while enabled); only the active locale loads; Workspace chunk
prefetched on a /workspace deep link
- Sagi: no work while disabled, pupil via CSS variables, memo avatar
- build: .br/.gz siblings, sourcemap hidden; service worker no longer caches
/assets/ (cache v3 drops the old one)
Tests: scripts/test-server.js drops the GIT_DIR-family variables a hook
exports, so a suite run from a linked worktree's pre-commit can no longer
commit into the real repository or set core.bare.
Docs: ARCHITECTURE, DATABASE, API, DEPLOYMENT, client README,
docs/plans/perf/README.md (audit, checklist, results).
AI Proxy (was AI Router):
- server/lib/ai-proxies.js, proxy-apply.js, routes/ai-proxies.js mounted at
/api/ai-proxies and the old /api/ai-routers; `router` still accepted on the
start routes; profiles move to <CLAUDE_HOME>/agenthub-proxies/
- dashboard_runs.router renamed in place to proxy: an older build on the same
DB fails on that column, so dev and prod sharing one DB update together
- a proxy's baseUrl stored as its root (a typed /v1 made /v1/v1/messages);
Test probes Claude /v1/messages and Codex /v1/responses
- Settings → AI Proxy (/settings/ai-router redirects), ProxyPicker
Updates:
- every tab reloads onto the new UI build when its WebSocket reconnects to a
server of another version (lib/newBuildReload.ts, once per version)
- Update now stops waiting on a plugin chain that failed or pulled nothing
(the old server never goes down), and reloads as soon as a new version answers
Also:
- a pane whose command exits at once says so (EPANEEXITED) instead of
"Command failed: tmux set-option … mouse on"
- the pane drag preview draws where the pane actually lands
Analytics parity phases 2–7 (docs/plans/analytics-parity/), built on
138e706 and ported onto v2.3.1:
- Cost by AI tool (?tool= on cost/analytics/stats/workflows, by_tool,
Codex fast/priority and flex tiers, agenthub_cost_usd_total{tool,model})
- Date ranges (from/to) on cost/analytics/workflows; Analytics range select
with Today / This week (Monday) / This month (1st) / 7d / 30d / 90d / All /
Custom; workflow filters carried into workflow-stats.js (read worker)
- Budgets and cost alerts as alert rule types (alert_rules rebuilt without
its rule_type CHECK), GET /api/budgets/status, rule preview, spend pace,
status-bar budget chip; rules stay editable in place
- Query Explorer (/query, allowlisted sort, CSV/JSON export) and scheduled
reports (report_definitions/report_runs, Settings → Reports)
- Session Replay tab and immutable snapshots (session_snapshots, /snapshot/:token)
- Remote Codex sync, push ingest with a hashed token, privacy redaction and
ignored folders (Settings → Data), Cursor sessions (no cost)
- Read-only MCP server, statusline, Grafana/Prometheus files
Also: resizable sidebar, pet run/wave animations, the workspace home keeps a
half-typed prompt, and the pre-commit hook runs only the related tests for a
small commit.
- GET /api/workflows: one LEAD() window pass in a read worker (read-pool.js,
workflow-stats.js); idx_events_session restored, idx_events_tool added
- compat-sqlite gains function/pluck/readonly so a plugin install without a
better-sqlite3 build boots again; antigravity-usage uses db.js's Database
- install-hooks adds no checkout hooks while the agenthub plugin is enabled
(events were doubled)
- terminal attach runs the live tmux server's own binary; the pane scrollbar
gets a visible 2-column style, and claude panes start with
CLAUDE_CODE_DISABLE_ALTERNATE_SCREEN=1 so tmux keeps history to scroll
- Codex sessions' conversation in GET /api/sessions/:id/transcript
- Codex/Antigravity sessions claimed by their workspace without hooks
- UI-audit follow-ups, docs and OpenAPI
- Workspace, SettingsHub, SessionDetail and the Workflow sub-tab are lazy
(main bundle 2,166 kB -> ~920 kB), wrapped in RouteErrorBoundary: a stale
chunk after a self-update reloads once (vite:preloadError too), any other
error shows a Reload panel instead of a blank page.
- Live refresh: lib/throttle.ts + hooks/useLiveRefresh.ts replace trailing
debounces that froze under a steady event stream; refresh skips hidden
tabs and refetches after a WebSocket reconnect. Monitor loads only while
shown, with one batched GET /api/agents?session_ids= instead of N+1; a
failing endpoint is an inline error, not a blank Dashboard.
- NeedsYouStrip on Monitor and above Analytics. Tabs, filters, page, sort
and search live in the URL (Dashboard, Sessions, Activity, SessionDetail).
- Sessions: tool / workspace / date-range filters, CSV/JSON export
(formula-neutralised), focusable row links, Open in workspace (the dead
/run links are gone), stale responses dropped, error + Retry.
- Fixes: "NaNs" durations, 30/7-day charts skipping idle days, weekday
labels west of UTC (now averages), cache hit rate denominator, billable
tokens vs cache reads, exact tool-call counts, "To" date = end of day,
invalid ?from= no longer throws, recent list no longer shrinks to 15,
stable message keys, transcript switch races.
- Light-theme-unreadable bg-black code panels and dark-only DAG fills use
tokens; status configs and chart colours use theme tokens; Health tab
and other literals translated.
- Server (additive): GET /api/agents?session_ids=, GET /api/sessions
awaiting=1&from=&to=, stats tool_call_count/distinct_tool_count, event
rows carry session/agent labels.
Analytics parity, phase 1 (data hygiene), from the analytics-parity session.
events.data is now stored through the SQL function trim_event_data(?)
(server/lib/event-payload.js, registered in server/db.js) on every event
INSERT and in the duplicate guard: Edit/MultiEdit/NotebookEdit/Write drop
the whole-file mirrors (originalFile, oldString/newString, Write content),
Read drops file.content/base64, and any string in tool_input/tool_response
over 8 KB keeps a 6 KB head + 2 KB tail. structuredPatch and every field
outside those two stay whole; _trimmed names what was dropped. The full
text stays in the transcript and its snapshot. Rows stored before this are
trimmed once in the background (sweepEventPayloads, 200 rows per write
transaction, user_version 4). No VACUUM, no row pruned.
The session list now prices every token column /api/pricing/cost/:id does
(speed, geo, tier, the 1h cache write, server-tool fees) through one
tokensForSessions helper, so a row's cost equals the session's own.
Right-click menus: one portal shell (components/ContextMenu.tsx) for every
item with actions - sidebar/Recents/archived session rows (SessionMenu),
console tabs, pane headers, the Changes panel's files/tree/commits, the
workspace home, sessions outside a workspace (SessionRowMenu: Sessions
table, Dashboard, Activity, Workflow runs), prompt images, every Settings
list and every Claude Code/Codex config list (FileItemMenu). Buttons stay;
a delete keeps its button's confirm; the shell stops portal clicks from
reaching the card that opened it.
Status bar: Workspace $ and Session $ on the Workspace page, from
?lane= on GET /api/pricing/cost and the focused/replayed session's cost.
Database, no data dropped: synchronous=NORMAL, journal_size_limit 64 MB,
hourly PASSIVE checkpoint + PRAGMA optimize, idx_events_session_created
(replacing the redundant idx_events_session), watchdog reads
sessions.transcript_path, transcript parsed before the hook's write
transaction, GET /api/sessions/:id?events=0 for the web page (reloads only
on its own session, debounced), streamed export, transactional clear-data
and chunked cleanup, subagent import reads tool_use_ids once, runtime logs
rotated past 5 MB.
token_usage holds a session's cumulative totals with no time, so
/api/pricing/cost booked a multi-day session's whole cost to its start
date and the status bar's Today read ~$0. TranscriptCache now also splits
usage by each message's UTC hour (tokensByHour); the hook writer stores it
in the new token_usage_hourly table (MAX per hour) and the server
back-fills existing sessions at start. The daily split books each total
over its hourly rows' local days and the rest to the start date, so the
days still add up to the total.
AI Router is a session's choice, not a workspace's:
- an AI Router select beside the model picker in the workspace home and a
pane's run form (RouterPicker.tsx; hidden off Claude/Codex or with only
Native; last pick remembered in agenthub.session.router)
- `router` on POST /api/lanes/:id/start and POST /api/run; Claude gets
--settings=<data dir>/router-settings/<name>.json (0600, outranks project
and user settings, key never in argv), Codex -p <name> + key via tmux -e
- dashboard_runs.router; a resume naming none goes back to it
- profile PATCH rewrites the session files, DELETE removes them
- gone: POST /api/lanes/:id/router/apply, LaneRouterRow, the Add-workspace
router radio, the claude-/codex-home `routers` answers; clearLaneRouters
takes the old keys out of each workspace's settings.local.json at start
Sidebar: Settings pinned to the foot, the collapse toggle beside it (stacked
on the rail) instead of in the brand's corner.
Icons: shipped SVG marks for claude, codex, openclaw and hermes (a
black/white pair swapped by theme) in client/public/ai-tools/, used by
AiToolIcon before the server's favicon.
Desktop-parity phase 6, following Claude Desktop:
- the workspace home's Worktree checkbox starts the pane with
`claude --worktree=<name>`, so Claude Code makes, populates
(.worktreeinclude, WorktreeCreate hook) and cleans up
<repo>/.claude/worktrees/<name>; AgentHub only excludes
/.claude/worktrees/ at the repo root (Claude does not)
- server/lib/session-worktree.js resolves a session's worktree from the
session row only (owned by the workspace, registered by git worktree
list); every lane git route takes ?session= and answers ENOWORKTREE
instead of falling back to the main folder
- the session view's Changes panel works in the opened run's worktree
(dashboard_runs.worktree, GitTarget in the client api)
- Resume/Fork start in the session's worktree; a gone one is ENOWORKTREE
- Archive of a worktree session removes the worktree after asking:
git worktree remove without --force, git branch -d (unmerged work keeps
its branch), a lock held by a live pid refused, a stale one cleared
- listWorktrees reads `locked <reason>` locks
- tests scrub GIT_* env so the pre-commit suite cannot act on this repo
Workspace sessions (docs/plans/desktop-parity, phases 1-5 and 7):
- sessions.lane_id: a workspace lists only the sessions its panes held,
stamped by bindRunSession (first writer wins, nested `claude -p`
excluded); GET /api/sessions?lane=&archived=
- session row state dot, double-click rename, archive (sessions.archived_at,
hides only); the activity bar's duplicate Sessions tab removed
- Fork beside Resume (--fork-session, Claude only), Ctrl/Cmd+click a live
session opens it in a second pane, a diff line's comment is pasted into
the workspace's last-focused pane (lib/panePaste.ts)
dashboard_runs stability:
- a pane's previous run is archived as <pane>@<started_at> instead of being
overwritten by INSERT OR REPLACE
- a run is flagged abandoned only once its tmux session is gone, at boot and
on the run-heal beat (panes outlive a server restart)
UI:
- sidebar: no footer, language or theme block; small collapse button in the
brand corner; version + update check in the connection details modal, a
status-bar dot while a release is out; language in Settings -> General
- one Switch component for every switch, the themed Checkbox for every
checkbox (no native checkboxes)
Also: AI Tools catalog trimmed to 10 built-ins (custom-file prompt modes
tested through fixtures), and the workspace plans under docs/plans/ (the
multi-user server plan lives in the AgentHub-Server repo).
/workspace?lane=<id> opens on the workspace's home (WorkspaceHome): a prompt
box whose Start opens a new session in the first free pane slot, Recents (a
live session attaches its pane, any other opens its transcript, nothing
resumes), and Instructions / Memory / Context / Folder. Instructions live in
lanes.instructions, extra folders and uploaded files in the new lane_sources
table (copies under <data dir>/workspace-context/<lane id>/); spawnRun reads
them from the lane's rows and hands Claude --add-dir + an appended system
prompt file, Codex --add-dir + developer_instructions. Memory is read-only.
Every dashboard pane carries AGENTHUB_RUN_ID; hooks forward it and the run
records which session the pane holds (a Claude pane follows /clear and
/resume but not a nested claude -p), so the sidebar and notifications open
the right pane (&run=). Settings loses its Projects group; Worktrees now
decides which worktrees Add folder suggests.
Pricing: the LiteLLM sync now adds every current Claude / GPT (>= 5) /
Gemini (>= 3) chat model before it is used and keeps its own rows
(model_pricing.source = 'sync', the Auto badge) at the published rate. A
user's row (any PUT) and a deleted pattern (model_pricing_dismissed) are
never touched; a deleted built-in row is not re-seeded at the next start; a
reset re-applies the list at once; retired Claude models are priced only
once used.
Models: GET /api/ai-tools/:id/models reads each tool's own list (Codex's
models_cache.json, Claude's aliases named after the newest version plus
priced versions, a catalog tool's vendor rows plus the models its sessions
used). A catalog `model: {flag, vendor?}` field passes `<flag> <id>`
(validated: never a flag or shell text); a per-tool default model starts
each run, and a resume keeps the session's own model.
Settings: every installed, enabled catalog tool without its own tab gets
one (AiToolConfig.tsx): path, homepage, Update, sessions/cost/lanes,
default model, vendor pricing, and the config files its release catalog
lists, edited over the opened sha, syntax-checked, backed up, written
atomically; credential-looking names refused. Antigravity keeps its hooks
card there.
Lanes: a just-started run no longer reads "dead" (age counts the run's
start while the old session is still bound); a second dashboard on the
same database re-broadcasts the lanes the other changed (PRAGMA
data_version, every 2 s); a lane start with a bad model answers 400.
DB: every migration runs under a cross-process lock (<db>-migrate.lock), so
two dashboards started at once on a fresh database both come up.
Part B: Model Pricing, Hook Configuration and CLAUDE_HOME move from
Settings → Dashboard to the Claude Code tab's AgentHub sub-tab
(ClaudeAgentHubSettings.tsx); CODEX_HOME is editable in Codex → AgentHub.
A home that is or contains the OS home is refused (EBADHOME).
Settings persistence: dashboard-saved settings go to
~/.config/agenthub/settings.env (AGENTHUB_SETTINGS_FILE), outside the plugin
cache an update wipes. Load order: real env > the checkout's .env > that file.
The first CLAUDE_HOME change pins DASHBOARD_DATA_DIR so the DB stays put.
C1: codex-usage.js reads Codex rollout journals into sessions (new column
tool) and token_usage, tracked per file in codex_rollouts; forks deduped by
total vector, sub-agent threads folded into the root; pricing-sync also
prices OpenAI models; PricingSettings with a vendor filter.
C2: codex-hooks.js installs AgentHub's hooks into $CODEX_HOME/hooks.json on
a click, through a stable forwarder; Codex trust is never written. The hook
route creates tool='codex' sessions and folds sub-agent hooks into the root.
D: antigravity-usage.js reads agy's per-conversation SQLite files (tokens
from ModelUsageStats) into tool='antigravity' sessions; antigravity-hooks.js
installs the named hook `agenthub` (PreInvocation, PostToolUse, Stop — never
PreToolUse) into ~/.gemini/config/hooks.json; Settings → Antigravity tab;
Gemini pricing.
Whole-change review fixes: sameOriginGuard on both home PUTs, routers.missing
on a CLAUDE_HOME change, reset-pricing scope, GET /api/sessions?tool=,
clear-data covering the new tables, Codex child-thread events moving no
status, hooks owning status only while current, EFOREIGN for a forwarder
that is not ours, agy quiet window and turn rules, AGENTHUB_SETTINGS_FILE
honoured from .env.
Docs, OpenAPI and the CLAUDE.md marker updated. Server 2141, client 779.
The (cwd, tool) rebuild lists its columns by hand and omitted
router_profile, which its probe adds earlier in the same start — a legacy
DB lost the column (and every lane's router) until the next restart.
- tmux targets are exact (`=name`, `=name:` for panes/options): a bare
`-t agenthub-lane-1` resolves by prefix to agenthub-lane-12 once lane 1
is gone, so killing a finished lane could kill another.
- getRun/killRun refuse ids that are not `agenthub-lane-<n>`, so
DELETE /api/run/main can no longer kill a user's own tmux session.
- router/apply reports restart_required only for a running pane; getRun
returns a `gone` handle, so it was always true.
- needs_action is cleared only by the session that raised it (new nullable
lanes.needs_action_session); a second session in the same cwd used to
clear it on its second hook.
- Changes panel at full screen: the whole Files view (finder, expand-all,
tree) becomes a left sidebar and the Files tab is dropped; a working-copy
file opens straight into the editor (Save/Cancel live once the text
differs, Cancel restores it, Whole file toggles back to the diff), and the
page is height-locked so the editor no longer collapses with no live run.
- Overview splits into Staged Changes / Changes groups in the work range.
- A changed file's whole tree row takes its status tint.
- Drag an image onto the lane terminal: POST /api/run/drop-image saves it
(png/jpeg/gif/webp, random name, 0600, pruned after a day) and the path
is pasted into the pane, which is how claude attaches it.
- Claude Opus 5.5 gets its own pricing row (4/20/0.2/5/8, fast 8/40);
claude-opus-5% had been pricing it as Opus 5.
- server/lib/pricing-sync.js: a model family used in token_usage with no
row of its own is inserted from LiteLLM's public price list, never
updated; off with AGENTHUB_PRICING_SYNC=0. repo-links allows exactly
that one third-party GitHub raw URL.
A doubled hook registration (plugin + leftover checkout install) POSTs every
event twice with a byte-identical payload, so the timeline showed each row
twice and token/alert/lane side effects ran twice. POST /api/hooks/event now
skips a payload identical to the session's LATEST event from the last 10s
(findRecentDuplicateEvent). Latest-only keeps two legitimately identical
Stop events of consecutive turns, which always have other events between.
Three defects that all showed up as "the dashboard's numbers are wrong", plus
the notification and brand work asked for alongside them.
Tokens: Claude Code writes ONE JSONL line per content block of an assistant
message, and every one of those lines repeats the same message-level `usage`.
Both parsers summed them, so a tool-heavy session's tokens — and its cost — ran
1.5-2x high (one 4h session: $45.82 stored against $27.72 real). `lastUsageMsgId`
counts a message once, carried across the transcript cache's incremental chunk
boundaries. `replaceTokenUsage`'s high-water baselines mean the fixed parser
cannot correct what is already stored, so `setTokenUsage` (exact, baselines
cleared) is added for the deliberate re-parse path and `reconcileTokens` now
covers every session with a transcript on disk rather than imported ones only.
Turns: every `TurnDuration` row was stamped NOW while its dedup guard matched on
the transcript's timestamp, so each hook re-inserted the session's whole turn
history — 587k rows, 75% of a real events table, one session holding 68,747
copies of the same turns. The row now carries the transcript timestamp, which
fixes the guard and the feed's ordering together. Same class of bug in the
session metadata: `thinking_blocks` / `turn_count` / `total_turn_duration_ms`
are assigned from the full parse instead of `+=`, since `extract()` returns
whole-transcript totals on every call.
Events: leftover checkout hook entries in `~/.claude/settings.json` double every
event, every subagent row and every token figure. `stripLegacyHooks` has always
removed them — but only inside `bootstrap()`, which `fastPathOk` skips on
exactly the healthy installs where nothing else forces a full pass, so they
survived indefinitely. `hasLegacyHooks` (pure fs) now blocks the fast path.
A one-time repair at the first start after upgrading (`user_version` 3) deletes
the duplicated `TurnDuration` rows and re-parses every transcript with exact
writes. Marker 3 and not 2: an install that ran the earlier pass had its
corrected rows re-inflated by whatever old build was still ingesting hooks.
Router: switching a lane's AI Router while a pane is live only half-applies —
Claude Code re-applies `env` to a running session but keeps the model it
resolved at startup, so the old router's model name reaches the new endpoint and
the next prompt fails with "model not found". The apply route reports
`restart_required` / `run_id`, and the picker offers a Restart that respawns the
pane from the lane id alone (replaying the old `--model` would re-create the
mismatch).
Notifications (#29): the two moments Claude is BLOCKED on the user now have
their own switches and default to on — `onAwaitingInput` (a permission prompt or
a question) and `onTurnComplete` (end of turn). Before, the first was unswitchable
and the second hung off a preference that defaults to off, so the one thing a
user walks away from went unannounced. Prefs shape/defaults/storage move to
`client/src/lib/notifPrefs.ts`; Settings and the hook had private copies.
Brand: the sidebar shows the SmartGift wordmark with the AgentHub caption under
it, the collapsed rail shows the mark, and the favicon is vendored to match.
Claude Opus 5 had no pricing row at all, so every token it burned was costed
at $0 — the pattern list stopped at the 4.x Opus family and nothing matched
`claude-opus-5`. Add it (5/25, fast mode 10/50, same shape as 4.8).
Sonnet 5's launch discount expired 2026-08-31 and the discounted rate became
the published standard, so the row's 3/15 standard rate started overcharging
the day after. Fold 2/10 into DEFAULT_PRICING and empty DEFAULT_INTRO_PRICING;
no model ships a promo today.
Fable/Mythos 5.1 fell through to the `claude-fable-5%` row, which caches at
$1/MTok against 5.1's $0.25 — give them their own rows. That overlap is only
resolvable if the lookup prefers the more specific pattern, and `matchPricing`
took `LIMIT 1` with no ORDER BY, i.e. whichever row was inserted first. Order
by pattern length descending, which is what docs/DATABASE.md already claimed.
(The nested REPLACE(model_pattern, '%', '%') it dropped was a no-op.)
Retire Claude 3.x, Opus 4/4.1 and Sonnet 4 from the defaults. The startup
top-up only ever ADDS missing rows, so removing them from DEFAULT_PRICING
would leave every existing install untouched — prune them once, guarded by
PRAGMA user_version (0 -> 1) so a model deliberately re-added in Settings
survives later restarts. The same one-shot migration corrects the existing
Sonnet 5 row. Only rates are deleted; token_usage is never touched, and
Reset Defaults still restores the full current set.
A new **AI Router** tab in Settings, next to Codex, stores named router
endpoints (9Router, OpenRouter, a self-hosted gateway) as profiles: a base
URL, an API key, a default model and an optional per-tier override. The tab
only stores them; a lane CHOOSES one in the Add-lane modal, via a Native /
AI Router radio, and that choice is written into the lane's own config.
The two CLIs are configured differently, and the difference is theirs:
- a claude lane gets eight `env` keys merged into its own
`.claude/settings.local.json` (`.local` because the value holds a key);
- a codex lane gets `$CODEX_HOME/<name>.config.toml` — the layer
`codex -p <name>` loads — because Codex has NO project-level config.
Its base `config.toml` is never touched.
Native is a built-in profile rather than a code branch, so "go back to
native" is the same write path. It blanks the two connection variables
instead of removing them (an absent key at project scope lets a USER-level
router apply) and removes the model names instead of writing them (writing
them would override the user's own global model choice).
Key material never leaves the server: profiles live one file each under
`<CLAUDE_HOME>/agenthub-routers/` (0700/0600), `GET /api/ai-routers`
reports only `hasKey`, and there is deliberately no route returning one
profile in full.
Three things this had to get right, each found by review and verified
against the real tools rather than assumed:
- `wire_api = "chat"` is rejected outright by codex-cli 0.149 ("no longer
supported"), so every codex lane on a router would have failed to start.
The layer writes "responses", and writes `env_key` only when the profile
actually has a key — naming an unset variable is itself a startup error.
- The key must never be TYPED into a pane. `tmux.sendCommand` used
`send-keys` with a `KEY='…' cmd` prefix, which the dashboard streams to
every attached browser and bash appends to history; it now takes over the
idle pane with `respawn-pane -k -e`, passing env through argv the way
`newSession` always has. Verified on tmux 3.4: multi-element argv arrives
intact and the pane renders no trace of the value.
- Attaching COPIES a profile, so editing or deleting one in Settings has to
reach the lanes already using it. An edit re-applies it (a rotated key
would otherwise never reach the lane still holding the revoked one); a
delete resets those lanes to native and removes the Codex layer file.
Both report per-lane failures, which the tab surfaces.
`.claude/settings.local.json` is added to the lane's git exclude, for both
native and router lanes: `agenthub lanes reset` runs `git clean -fd` without
`-x`, so an un-excluded settings file is deleted on reset — and untracked
meant the key was committable.
DB: `lanes.router_profile`, nullable, additive. Written only by
`setRouterProfile` beside the disk write, never through `PATCHABLE`.
Lane Pipeline Pass Rates read `lanes.stages`, which is a snapshot of the
pass in flight: re-entering a node overwrites its record, and moving
backwards deletes every record ahead of the new stage. The report treated
one lane row as one run, so shipping ten features through a lane still
counted as one, and no node could ever exceed 1/1. History survived only
via `lane_features`, which is written when a feature is activated — a
path most users never take. The panel was not wrong, it was structurally
incapable of counting.
Add `lane_stage_events`, an append-only log written by every declaration
and every accepted detection, and aggregate from that. A run is one pass
through the template: the log starts a new one when the lane moves
backwards or leaves the template's last node. Existing snapshots are
seeded as one run each on first boot, so upgrading does not read as
"never run". Templates a lane is configured with are still listed at
zero — a pipeline going unused has to be visible as such.
Count inferred reaches too, in their own column and their own colour.
Detection stays out of the pass rate ("inference is never evidence"), but
reporting a stage the hooks demonstrably saw as "never reached" is what
made the panel look broken to anyone who works without running the
pipeline commands.
Also reject a stage naming no node of the lane's pipeline (400 EBADSTAGE,
listing the ones that exist) instead of storing it with a warning nobody
reads. A real lane had accumulated five such keys, one of them `test` on
a ship-feature lane whose CI node is `gates` — which is why that row read
never-reached while the stage had in fact been declared. Aliases still
resolve, so shorthand is unaffected.
Two leftovers from the rebrand, done together because they touch the same
docs and manifests.
**Data directory.** `agent-dashboard` could not simply be renamed in the
source: it names a real directory holding the live SQLite database. Startup
now moves it once, via `migrateLegacyDataDir()` in `server/lib/claude-home.js`
— a same-filesystem rename, so it is instant even at multiple gigabytes and
cannot leave two divergent copies. `getDataDir()` stays PURE and resolves to
the legacy path until the move happens; the first version of this put the
rename inside the getter, which meant `npm run test:server` relocated the
developer's real database as a side effect. The call sits behind
`require.main === module` and ahead of the router requires, because those pull
in db.js, which opens SQLite at module load. A failed rename is silent and the
old directory keeps being used: an unfashionable directory name beats a blank
database beside an unreachable one. Same treatment for the port-discovery file
(`.agent-dashboard.json` → `.agenthub.json`, old one dropped on write).
Not renamed on purpose: the Sagi localStorage keys (renaming silently resets
everyone's prefs) and the PWA manifest `id` (changing it orphans an installed
app).
**MCP server.** 4 tool calls in 46 days across 577,132 events, against 4,556
lines of TypeScript, a 2 MB committed build artifact, and a pre-commit guard
that had already blocked commits twice. Deleted: `mcp/`, the maintainer skills
and `mcp-reviewer` agent, the build check, `plugins/agenthub/.mcp.json`,
`docs/MCP.md`, 12 npm scripts, 5 Makefile targets, and every doc section
describing it.
Kept, because they are unrelated things sharing the name: the Config
Explorer's MCP tab and its API, per-lane Playwright MCP, the agenthub-config
plugin's MCP audit, and the historical `mcp__plugin_agenthub_*` events that
analytics must keep attributing.
Verified: server 1432/1432, client 459/459, tsc clean, prettier clean. Both
directions of the migration proven — a sandbox-HOME test run leaves the legacy
directory untouched, and a real boot moves it with its contents intact.
Clean break, no compatibility alias. Renames the product name, the CLI
binary (`ccam` -> `agenthub`), the env var prefix (`CCAM_` ->
`AGENTHUB_`), the Prometheus metric namespace (`ccam_*` ->
`agenthub_*`), the tmux session prefix (`ccam-lane-<id>` ->
`agenthub-lane-<id>`), the per-repo runtime profile directory
(`.ccam/profile` -> `.agenthub/profile`), the lanes state root
(`~/.claude/ccam-lanes` -> `~/.claude/agenthub-lanes`), the bundled
plugins (`plugins/ccam*` -> `plugins/agenthub*`) and their slash
commands (`/ccam-doctor` -> `/agenthub-doctor`, `/ccam:*` ->
`/agenthub:*`).
Because there is no alias, an existing install has to be updated by
hand: rename its `.ccam/profile` directory and any `CCAM_*` variables,
and reinstall the plugin under its new name.
Two other name families are deliberately left alone here, since they are
not just strings: the `Claude-Code-Monitor` repository/marketplace name,
and the `agent-dashboard` npm package plus its `~/.claude/agent-dashboard`
runtime directory, which holds the live database and needs a migration
rather than a rename.
Verified: server 1406/1406, client 459/459, `grep -i ccam` clean across
tracked files, `openapi.yaml` regenerated (95 paths, unchanged count).
Repo/Adopt mode (POST /api/lanes/ensure) treated lanes.cwd as globally
unique, so adopting a path already driven by one CLI silently ignored a
different tool choice. Uniqueness moves to the composite (cwd, tool):
one directory can now back a claude lane and a codex lane side by side.
- server/db.js: drop the column-level UNIQUE on cwd, migrate existing
DBs via table rebuild, add idx_lanes_cwd_tool.
- resolveLaneByCwd gains an optional tool filter; hook ingestion always
pins to "claude" since Codex has no hook mechanism here, so a hook
event never touches a codex lane sharing the path.
- POST /ensure resolves the existing owner by (cwd, tool) instead of
cwd alone, so a different tool creates a sibling lane.
- LaneConsolePane's cwd-only lane lookups now also match tool, so a
shared cwd doesn't attach a run to the wrong sibling lane.
Worktree mode is unaffected (already allocates a unique directory per
lane). docs/LANES.md and CLAUDE.md updated to describe the new
composite uniqueness.
Gives each lane its own slot-derived runtime (ports, detached process
lifecycle, profile-driven hooks) and its own database/Redis logical
index/.env file, so two lanes running the same repo's stack at once no
longer share state. Machine-level DB/Redis credentials live at
~/.ccam/secrets.env (mode 0600, never returned by any route); a hook's
output is redacted of that password (raw and URL-encoded forms) before
it reaches a log file or the lane_hook_output websocket broadcast.
Wired into provision/up/reset/remove; reset accepts --keep-db to skip
the drop/recreate/migrate/reseed block entirely.
Internal SmartGift build of a Claude Code monitoring dashboard.
Lanes: a durable unit of parallel agent work, one per working directory,
tracked across session restarts. Managed lanes are git worktrees the
dashboard provisions and can reset or remove behind a three-check destroy
guard and a counted preflight; adopted lanes are directories you already
own and are never destroyable.
Pipelines: a lane moves through pipeline stages. A stage the agent declares
with evidence renders green; a stage inferred from the tool-event stream
renders dashed amber and never counts as done. Detection is forward-only
within a 30-minute window, and never writes the declared stage.
Workspace: one page at /run with a lane grid, the selected lane's pipeline,
and a full Claude console behind a disclosure.
nntrivi2001