* feat(web): feature-flagged rich composer for inline session @ mentions
Custom segmented contenteditable (not TipTap) inserts caret-local session
atoms from the existing @ picker and serializes to markdown links on send.
Textarea path remains default until flag parity dogfood.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(web): rich composer mention boundary + #1215 refs
Treat U+FFFC mirror atoms as word boundaries so @ after a session
token still opens autocomplete. Point comments at Fixes#1215.
Co-authored-by: Cursor <cursoragent@cursor.com>
* test(web): peer-stack e2e for rich composer session @ mentions (#1215)
Smoke: flag on, @ picker inserts inline session atom chip (not prose dump).
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(web): preserve newlines in rich composer Enter-newline mode
Chromium splits contenteditable on Enter into block divs; serialize those
as \\n and insert <br> when parent leaves Enter unhandled (Shift+Enter /
enter-inserts-newline).
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(web): show @ badge when rich composer mentions flag is on
Dogfood was invisible: flag-off looks like a normal textarea, and flag-on
had no chrome. Surface a small @ badge when enabled.
* fix(web): rich session composer on by default (not a user setting)
The plan dual-path was an engineering kill-switch, not an opt-in. Default
to the segmented composer; only richMentions=0 disables. Drop the flag
badge and record a peer-stack motion proof covering chips + baseline UX.
* fix(web): make rich composer Shift+Enter create a visible newline
Trailing <br>+empty text node was a silent no-op at EOL. Use
insertLineBreak (ZWSP pad fallback), assert real \\n in peer e2e.
* feat(web): hover tooltips on rich composer session chips
Show full title, status, short id, and path on chip hover via a portal
bubble fed by live useSessions lookup (drafts fall back to title + id).
* fix(web): dismiss rich composer chip tooltips on mouse leave
contenteditable pointerout/relatedTarget was flaky so tips stuck after
leaving the chip. Hit-test on pointermove, clear on prose/input/leave.
* fix(web): address cold-review Blocker/Majors on rich composer
Exclude peer e2e from default Playwright; force plain-text paste; restore
newline hard-stop in findActiveWord; fix root-anchored selection mapping
and nested-block serialize; cover with unit tests.
* chore: drop accidental .cursor files from rich-composer tip
* fix(web): close remaining cold-review gaps on rich composer
Drop absolute peer e2e tooling imports, prove chip→markdown send, and
harden paste/EOL/focus/tooltip/Enter edges before Meta rematerialize.
Co-authored-by: Cursor <cursoragent@cursor.com>
* chore: absorb soup playwright.config union for clean remat
Keep fork peer-stack timeouts/annotated-video wiring and add testIgnore
for e2e/peer so the next driver rematerialize does not conflict.
Co-authored-by: Cursor <cursoragent@cursor.com>
* revert: drop fork playwright tooling from upstreamable tip
Peer-stack annotated-video + HAPI_PEER wiring stay on fork main / soup.
Product tip only needs testIgnore for e2e/peer (see docs/tooling/peer-stack.md).
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(web): fix rich composer Shift+Enter double newline and paste space
Prefer manual newline+pad over execCommand insertLineBreak, and stop
applying autocomplete trailing-space on paste/drop paths.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(web): pad EOL Shift+Enter after Range.insertNode split
insertNode always leaves an empty text sibling, so !nextSibling never
saw EOL; detect meaningful trailing content and cover with jsdom tests.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(web): drop custom onDrop from rich composer
Intercepting drop without caretRangeFromPoint landed text at EOF or
no-oped in-editor moves. Native CE drop is enough for #1215; paste
still forces plain text.
Co-authored-by: Cursor <cursoragent@cursor.com>
* feat(web): sidebar-parity tooltips on rich composer session chips
Reuse SessionRowSummary (flavor, thinking/attention, schedule, todos,
relative ago, path) for chip hover so the tip matches the session list.
Co-authored-by: Cursor <cursoragent@cursor.com>
* chore: keep peer-stack e2e off the upstreamable tip
Peer specs and playwright.peer.config stay on fork main per
docs/tooling/peer-stack.md; default config still testIgnore's e2e/peer.
Co-authored-by: Cursor <cursoragent@cursor.com>
* feat: cite sessions with UUID wire + inspect_peer for agent/overseer
Rich composer chips already serialize to [title](/sessions/<id>); flush
before send so the agent prompt never gets title-only chip text. Add
inspect_peer (MCP + hapi inspect-peer) as the read twin of ping_peer so
that same id is immediately usable for overseer/agent peer lookup, with
system-prompt glue from citations to inspect/ping.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): gate inspect_peer behind permission approval
Cross-session history reads need the same prompt path as ping_peer:
keep inspect_peer off Claude --allowedTools and treat it as sensitive
in ACP/OpenCode read-only mode so prompt injection cannot silently
enumerate peer transcripts.
Co-authored-by: Cursor <cursoragent@cursor.com>
* chore: clarify playwright peer testIgnore is upstream-safe
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(web): keep session UUIDs on rich composer copy/cut/paste
Copy/cut write wire markdown so chips do not collapse to @title-only
clipboard text; paste reparses session links back into atoms.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
Follow-up to #1268. The per-file `mkdtempSync`/`join(tmpdir(), ...)` HAPI
homes in the cursor-models test cluster were never removed, so every run
leaked a directory into the system temp dir (afterEach only cleared the
cache file inside them). Save/restore HAPI_HOME and recursively remove
each per-file temp root in teardown across all three cursor-models test
files (cursorModels, cursorModelsSharedCache, and the stale-lock test that
had the same pattern). No source/runtime change.
Co-authored-by: Cursor <cursoragent@cursor.com>
The four cursorModels* CLI test files share one on-disk cache path
($HAPI_HOME/cache/cursor-models.json, defaulting to /tmp/hapi when
HAPI_HOME is unset). Two files already isolate it (cursorModelsStaleLock
via a PID-namespaced home; handlers/cursorModels via a unique temp home),
but cursorModels.test.ts and cursorModelsSharedCache.test.ts do not.
Under vitest's parallel file execution, cursorModelsSharedCache's
afterEach(_resetSharedCursorModelsCacheForTests) rmSyncs that shared file
between the other file's write and read, so the read returns null and
"inherits cliModelSkus from shared cache" fails with
`expected undefined to deeply equal [...]`. Passes in isolation; fails at
random in the full parallel suite.
Give both un-isolated files their own mkdtemp HAPI_HOME at module load so
each test file's cache path is unique regardless of worker-process reuse.
Fixesheavygee/hapi#101
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(web): exclude subagent usage from the parent context indicator
The status bar's `ctx N/M` and `cache N` come from latestUsage, which
scans the normalized messages backwards for the most recent usage. That
scan includes sidechain messages, so while a Task subagent runs its
usage — describing the subagent's own, much smaller context — becomes
the parent's numerator, then snaps back when the parent resumes.
The existing `scope_role !== 'child'` guard never fired on any path.
Claude never stamps scope_role (sdkToLogConverter.ts says so outright),
and Codex drops child token_count events in the CLI before they can
reach the web layer, so no producer ever emits 'child'. isSidechain is
the signal that actually survives.
sdkToLogConverter.ts:308-313 already documents this exact reducer
behaviour, but works around only the denominator by forcing the main
session's context_window onto sidechain messages. The numerator was
left unguarded.
* fix(cli): stop stripping context_window from local-session usage
UsageSchema is a plain z.object, so Zod's default strip mode drops every
undeclared key. sessionScanner forwards parsed.data rather than the raw
line, so on the local-JSONL path usage is truncated to the five declared
fields and context_window — injected on the SDK path by
sdkToLogConverter — never survives.
The web status bar then falls back to getContextBudgetTokens, which
subtracts a 10k headroom, so the same model reports a 1.0M denominator
on a remote session and 990k on a local one.
RawMessageSchema right below already carries .passthrough() with a
comment about losing message.model and messageId the same way; the
nested usage object just never got the same treatment.
서브에이전트(sidechain)가 오래 걸리는 도구를 실행할 때 SDK가 주기적으로
내보내는 tool_progress heartbeat 이벤트가 isClaudeChatVisibleMessage()의
기본 통과 분기를 거쳐 raw JSON 그대로 채팅에 노출되던 문제를 고친다.
rate_limit_event 필터링(#423)과 동일한 패턴으로 타입 전체를 deny한다.
* fix(cli): surface real Cursor ACP session/load errors
Stop mislabeling every ACP session/load failure as a legacy
stream-json protocol problem. Prefer Cursor's Cannot use this model
stderr (including Available models when present), attach drained
stderr on process close, and keep structured formatAcpLoadError logs.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): accumulate ACP stderr across split chunks
Address Codex Major on #1198: child_process stderr data events are not
message boundaries. Concatenate raw chunks in a rolling window, extract
Cannot use this model from the window on close, and prefer that over a
partial onStderrError hint when classifying resume failures.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): pin ACP model-rejection stderr when catalog overflows
Once Cannot use this model appears, keep the buffer from that match
head so a long Available models list cannot roll the rejection out of
the rolling window (Codex follow-up on #1198).
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): wait for model id before ACP model-rejection emit
Only emit Cannot use this model via onStderrError once a non-space
token follows the colon, so a split before the id cannot suppress the
completed rolling-window message (Codex Minor on #1198).
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): block ACP writes between process exit and close
Keep the post-exit stdin write guard while deferring markClosed until
stdio close so stderr can still enrich the failure (Codex Minor on #1198).
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
Claude Code injects its own user-role turns for skill bodies and compact
continuation summaries. The on-disk transcript flags them `isMeta`, which
claudeLocalLauncher drops before they ever reach the web UI. Over
stream-json the same event is flagged `isSynthetic` instead, and
sdkToLogConverter copied only `message`, dropping the flag entirely.
With no `isMeta` on the converted line, every downstream guard let it
through: OutgoingMessageQueue forwarded it, isExternalUserMessage
classified it as genuine human input (its XML-prefix allowlist does not
match a bare-markdown skill body), and the web UI rendered the full skill
document as a user bubble.
Normalize `isSynthetic` to `isMeta` in the converter so the SDK path
carries the same signal as the transcript path and the existing filters
fire. Fixes skill injections appearing as user messages in remote mode.
* fix(cli): restore Pi session resume
Use Pi's supported --session flag and keep the session initialized by the CLI instead of replacing it with a racing new_session RPC.
* test(cli): cover fresh Pi startup
* fix(web): expose Codex Fast and Plan on Create Session
Wire serviceTier and collaborationMode through spawn so Create can set
the same Codex options chat Settings already supports (#1015).
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): forward collaborationMode through machine spawn RPC
Create Session Plan was accepted by the hub but dropped in apiMachine
before buildCliArgs; also preserve collaborationMode on resume spawn.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): correct stopSession mock type in spawn RPC test
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(web): keep Fast mode across Create draft restore while models load
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(web): preserve pending Fast selection
* fix: apply Fast and Plan to imported Codex sessions
* test: narrow imported Codex session id
* fix: forward explicit Standard service tier
* fix: integrate create-session controls with current main
* test: close Codex RPC suite
* fix: preserve existing session spawn field
* fix(web): integrate Codex controls with current New Session form
* fix(web): reconcile draft types and submit state
* fix(hub): integrate spawn arguments with current resume flow
* test(cli): isolate spawn RPC suite
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
* feat(cli): add ping-peer CLI and MCP ping_peer for peer messaging
Promote resume-if-inactive + wait-active + POST message into a first-class
CLI command and session MCP tool so agents stop reinventing JWT+curl.
Fixes#1194
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): do not auto-approve MCP ping_peer
Cross-session messaging can resume a peer and inject a prompt, so keep
permission-mode gating (Codex PR review Major on #1195).
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): require approval for ping_peer in read-only mode
Read-only auto-approve treated non-write names as safe; ping_peer can still
resume a peer and inject prompts, so gate it like a write tool.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): keep ping_peer out of Claude --allowedTools
toolNames still registers the MCP tool, but Claude auto-allow must not
pre-approve cross-session resume+inject without a permission prompt.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): re-check session active before ping-peer send
List/get can race; POST /messages still 409s if the target flips inactive
before send. Resume+wait again (and re-gate pi) immediately before POST.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
- prepare-npm-packages: exit(1) instead of warn-and-continue when a
platform binary is missing, so a broken build aborts the release
- release-all: after publishing platform packages, poll npm view for
every @twsxtd/hapi-<platform> package until it matches the release
version (10min timeout, 15s interval) before publishing the main
package
Fixes#1149
* fix(cursor): nest ACP extension outcome so plan approvals aren't cancelled
Cursor's ACP blocking extension methods (cursor/ask_question,
cursor/create_plan) expect the JSON-RPC result to nest the outcome under
an `outcome` key, e.g. `{ outcome: { outcome: "accepted" } }`. The
adapter returned it flat (`{ outcome: "accepted" }`), so Cursor read
`response.outcome.outcome` as undefined and fell back to a cancellation
— an approved plan was relayed to the agent as `User cancelled`, making
plan mode unusable over HAPI for Cursor sessions.
Wrap every ask_question / create_plan response in the nested envelope
and add regression tests asserting the exact wire shape for the
affirmative approval -> CreatePlan path (plus approved_for_session,
reject, and abort).
Fixes#79
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cursor): continue task after CreatePlan Yes (plan→execute)
Nested ACP accept alone is not a complete fix: Yes unblocked create_plan
but the prompt turn still ended with the plan "done" and no execution.
Mirror Claude ExitPlanMode: on accept, leave plan/ask for an executable
mode and queue a continue prompt so the original user task keeps going.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cursor): sync runCursor enqueue mode after CreatePlan accept
Codex Major on #1097: setPermissionMode alone left runCursor's
currentPermissionMode stale, so the next user message could re-enter
plan/ask after Yes. Notify onPermissionModeChanged from CursorSession
so the enqueue source of truth stays aligned with the session.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Debian <heavygee@oos-linux.in.lockhouse>
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): drop unknown SDK message types instead of passing them through
The SDK-to-log converter's switch had a fail-open default that stamped any
unrecognized SDK message with transcript base fields (parentUuid/sessionId/
userType) and forwarded it. Claude Code emits a tool_progress heartbeat every
30s for long-running tools, so a single slow Bash call flooded the chat: the
web normalizer matches no known shape for those records and falls back to
rendering the raw envelope as message text.
Gate the switch on an explicit allowlist instead, bailing before the uuid is
allocated so a dropped event cannot advance sidechain/parent tracking -- the
heartbeats share one parent_tool_use_id and were overwriting the pointer a
subagent's next real message parents to. This matches the local launcher,
which already enforces the same allowlist via RawJSONLinesSchema.safeParse.
The default branch stays as a fail-closed guard so adding a type to the
allowlist without a matching case drops the message rather than leaking it.
* fix(cli): re-check reassembled text for internal event JSON at flush boundary
isInternalEventJson was only applied per incoming chunk. In delta mode
(OpenCode) every chunk is a fragment, so none of them parses as JSON on its
own and the filter never fires; the pieces accumulate and flushText emits the
reassembled envelope verbatim. The dedupe path has the same hole whenever two
chunks share no overlap.
Check again in flushText, which is the first point the complete text exists,
and tolerate surrounding whitespace so an envelope preceded by a newline is
not waved through by the leading-'{' fast path.
Genuine assistant output that happens to be JSON is unaffected: the matcher
still requires the specific { type: 'output', data: { parentUuid, sessionId,
userType } } envelope shape.
* fix(cli): fail closed on unrecognized agent message in converter
convertAgentMessage's exhaustiveness default returned the message object
itself at runtime. The never binding makes the branch unreachable today, but
every caller forwards a non-null result straight into the chat stream, so the
failure mode if it were ever reached is a raw object on screen. Keep the
compile-time check, return null at runtime.
* test(cli): cover command_lifecycle, a second unknown type seen leaking
Observed in the same session after tool_progress. The allowlist already
covered it with no code change, which is the argument for gating on known
types rather than adding a case per offender.
* feat(cli): export HAPI_SESSION_ID into wrapped agent env
Publish the hub session id into process.env at session bootstrap so every
downstream agent spawn inherits it. HAPI runs one hub session per CLI process
(the runner forks a fresh hapi child per session; local is 1:1) and every
flavor's agent spawn derives its child env from process.env, so a single seam
covers claude / codex / cursor / gemini / opencode / kimi / grok / pi -
runner-spawned and local - plus future flavors, without touching each launcher.
Agents can read HAPI_SESSION_ID to self-target their own hub session over REST
or shell helpers without listing /api/sessions. Prefer the MCP display_image
tool for inline media when available; HAPI_SESSION_ID is the deterministic
fallback for non-MCP tooling.
Closes#1119
Co-authored-by: Cursor <cursoragent@cursor.com>
* feat(scripts): self-target hapi-display-image via HAPI_SESSION_ID
Teach the in-tree shell helper to use $HAPI_SESSION_ID for path-only /
self invocations: GET /api/sessions/:id directly instead of listing
/api/sessions. Explicit session prefixes keep the previous list path.
Gives #1119 a tangible now benefit - the tool that forced the wasteful
list-and-reverse-lookup dance no longer needs it inside a wrapped session.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): defer HAPI_SESSION_ID export until lazy Codex materializes
The provisional lazy-session id was exported at bootstrap before the hub
row existed, so path-only self-targeting (GET /api/sessions/:id) could
404 while materialization was still pending. Export on onMaterialized
instead, and await materialize in buildHapiMcpBridge before starting the
MCP server / spawning Codex so the agent inherits an id the hub can
resolve (and so hapiMcpUrl is persisted, not only local pending state).
Addresses Codex review Major on #1121.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Debian <heavygee@oos-linux.in.lockhouse>
Co-authored-by: Cursor <cursoragent@cursor.com>
Hub already passes access.sessionId on resume (#1088); Cursor CLI still ignored
it (Codex-only). Parse/pass the flag for cursor and lock in reuse-without-ready-wait tests.
Co-authored-by: Debian <heavygee@oos-linux.in.lockhouse>
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): buffer Pi prompts until RPC startup ready
A prompt POSTed immediately after spawn (a supported handoff pattern used
by hapi-ping-peer and intake scripts) could reach `pi --mode rpc` before
its `new_session`/`get_state` startup finished, wedging the turn:
`agent_start` then silence, no tool calls. The socket goes `active` (spawn
success) well before Pi's session is initialized, so `active` is not a
safe ready signal for Pi.
Gate outbound prompt/steer sends behind a startup ready gate on PiSession:
`runWhenReady()` delivers immediately once ready, else buffers FIFO;
`markReady()` fires on the first `get_state` response (the signal that
persists `metadata.piSessionId`, which working callers already wait for)
and drains the buffer in order. A 30s unref'd fallback timer force-drains
if `get_state` never lands, degrading to prior send-anyway behaviour
rather than swallowing the message forever.
Fixes#1143
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): honor cancel-queued-message for buffered Pi prompts
Addresses the MAJOR review finding on the startup ready-buffer: while a
prompt is held behind runWhenReady, the hub can send cancel-queued-message
for its localId. Pi registered no onCancelQueuedMessage handler, so
ApiSessionClient acked removed:false, the hub marked the row invoked, yet
the buffered closure still drained on get_state and fired the cancelled
prompt.
Carry the localId with each buffered send and add
PiSession.cancelBufferedMessage, then register apiSession.onCancelQueuedMessage
so a cancel drops the still-buffered prompt (returns true) instead of
sending it. Once drained to Pi it cannot be recalled — returns false,
matching the other agents' queue.cancelByLocalId best-effort semantics.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
Cursor ACP (and other remotes) flagged the glued-on SKILL_LOOKUP_INSTRUCTION
as prompt injection. Keep discovery on the skill_lookup MCP tool description
and on system prompts (OpenCode/Grok); do not taint user messages.
Fixes#1095
Co-authored-by: Debian <heavygee@oos-linux.in.lockhouse>
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(cli): cap and back off consecutive remote launch failures
claudeRemoteLauncher's respawn loop retried claudeRemote() immediately
on every throw with no backoff or limit. A deterministic launch
failure (bad auth, invalid model/args, spawn failure) respawned in a
tight loop instead of giving up, hammering the same failure forever.
Track whether onReady() fired at least once per attempt to tell an
immediate/deterministic failure apart from a failure after real
progress, back off between immediate-failure retries, and after 3
consecutive immediate failures drop the message that keeps triggering
them and reset the streak, instead of respawning forever. The session
keeps running so a later, unrelated message still gets its own budget.
The streak reset on a non-throwing attempt is itself gated on having
reached onReady, not applied unconditionally -- otherwise a message
that keeps getting parked and re-picked-up on alternating attempts
(e.g. an isolated command hitting the same deterministic failure)
would reset the streak every other attempt and the cap would never
fire.
* fix(cli): restore queued message when remote launch fails before delivery
MessageQueue2.collectBatch() acks a message (fires onBatchConsumed,
which the hub uses to mark it consumed) at dequeue time, before the
message ever reaches the SDK. If claudeRemote() then throws before
onReady -- e.g. the process dies right after picking up the message --
the catch block only logged and retried, so the message vanished: the
hub already thinks it was delivered, but the CLI never acted on it.
Track the message returned from nextMessage() (whether freshly
dequeued or held in `pending` across a mode change) as in-flight until
the next onReady confirms it was handled, and restore it to the front
of the queue (preserving isolation via unshiftIsolated when needed) if
the attempt throws and will be retried. Restoring happens even if the
throw races with a user-initiated switch/exit, so a message is not
silently dropped by that unrelated shutdown either.
When the immediate-failure cap from the previous commit is reached,
the in-flight message is dropped instead of restored: unshifting it
back would just feed it into another immediate failure on the very
next attempt, storming again. This mirrors
cursorLegacyRemoteLauncher's existing drop-and-reset policy on its own
consecutive-failure cap.
* fix(cli): preserve localId when restoring a failed message batch
MessageQueue2.collectBatch() already collects each queue item's
localId (it fires onBatchConsumed with the full list to ack them), but
only exposed the joined `message` string to callers, discarding the
per-item localIds and their original boundaries in the process.
When claudeRemoteLauncher restores a dequeued-but-undelivered batch
after a launch failure, it re-added the joined string as a single new
queue item with no localId, orphaning the retried prompt from the hub
row(s) it originated from (and from cancel-by-localId).
Expose the pre-join `items` breakdown (message + localId per item)
alongside the existing joined `message` field on
collectBatch()/waitForMessagesAndGetAsString() -- purely additive, so
the other callers of waitForMessagesAndGetAsString() (grok, kimi,
opencode, cursor, codex, runAgentSession) are unaffected. On restore,
unshift each original item individually in reverse order, so the
localId and relative order of a multi-message batch are both
preserved instead of just the first item's.
* fix(cli): reset immediate-failure streak on a delivered non-onReady success
claudeRemote.ts's /clear handling delivers the queued message to the
SDK, then calls onSessionReset()/onCompletionEvent() and returns
successfully without ever calling onReady(). The success-path streak
reset only cleared on reachedReadyThisAttempt, so a successful /clear
between two unrelated immediate launch failures did not reset the
streak: an unrelated message's very next failure could hit the
3-in-a-row cap after just 1 failure, and the resulting banner would
misreport "3 times in a row".
Track whether nextMessage() actually handed a message to the SDK this
attempt (deliveredMessageThisAttempt), separately from whether the
attempt reached onReady, and reset the streak on either signal. The
livelock-prone case this guards against (a message parked into
`pending` and the attempt returning without ever delivering anything)
leaves both flags false, so it still does not reset the streak.
* feat(shared,cli): whitelist away_summary so auto recap reaches the hub
Claude Code's local TUI writes an automatic away-summary recap to the
session transcript on window blur/focus (5min+ idle), but
VISIBLE_CLAUDE_SYSTEM_SUBTYPES dropped it before it ever reached the
hub. Add it to the whitelist so the local launcher forwards it like
the other system subtypes, and cover the forwarding + Zod passthrough
of the recap `content` field with tests.
* feat(web): render Claude Code's automatic away recap in the chat
Once away_summary reaches the hub (previous commit), the web chat
still dropped it silently: normalizeAgent had no branch for the
subtype, so it fell through to `return null`. Add a `recap` AgentEvent,
a normalizeAgent branch mirroring the existing turn_duration/compact
subtype branches, and a presentation entry that prefixes the text with
`recap:` so it reads distinctly from the manual /recap assistant
bubble (which already renders as a normal message). No new render
component needed: it flows through the existing generic system-event
row (SystemMessage.tsx + getEventPresentation) that every other system
subtype already uses.
* fix(web): drop inaccurate manual-/recap comparison from recap comments
* perf(claude): scan transcripts incrementally
The claude session scanner re-read the entire transcript JSONL on every
scan, so the cost of each poll grew with the length of the conversation.
Track a byte offset per file instead and parse only the bytes appended
since the previous scan.
A trailing partial line — a write still in progress — is held back until
its newline arrives. A file that shrank resets the cursor to 0; the base
scanner's uuid dedup absorbs the re-sent events. A read that fails
returns no events and leaves the cursor where it was, so a transient
error is retried on the next scan rather than skipping content.
The codex scanner received this in #1031; this extends the same
improvement to the claude scanner. readSessionLog is exported for tests,
mirroring readTranscriptRange there.
* fix(claude): forward a complete final record with no trailing newline
The incremental reader consumed only through the last newline, so a final
JSONL record flushed without its terminating newline — at shutdown or on
import — was held back as if it were a partial write and never forwarded
until a later append supplied the newline. The previous whole-file reader
parsed such a record.
Consume a trailing segment when it already parses as a complete JSON value,
and keep holding back a genuinely partial line (which parses as incomplete).
