The sheet is sized to the transcript, so it changed height on every reply with
no transition — the panel twitched to each new size instead of growing into it.
Animated on the reveal timing.
The edge flip was reading the window, which stopped being the same thing as the
HUD once the sheet started sizing itself: a tall window's top edge reaches the
screen top while the visible bar is still well down the display, so it flipped
far too early. It now measures the visible panel, against a threshold
proportional to the display rather than an exact-flush rule — the HUD hugs its
bar, so its visible top can never actually touch the screen edge and a
zero-tolerance test would never fire at all.
Parked at the top the whole gradient flipped with the layout, which inverted the
thing it exists to do. The transcript reads oldest-to-newest downward wherever
the bar is, so the newest turn is at the bottom either way — flipping the mask
faded the newest reply and left the oldest solid, leaving a stray bubble hanging
over the desktop with the reply dissolved beneath it.
Only the anchor flips now: the sheet hangs from the bar instead of standing on
it, the thread's padding moves, and the exit chip drops clear. The fade always
runs away from the newest message.
Four things the band's content-sizing exposed, all of them stale assumptions
from when it was a full-window box.
A leftover `:focus-within` rule still painted the band's own background, so
focusing an empty session filled the whole window with a solid slab even though
the sheet had correctly collapsed to nothing.
The exit chip anchored to the band rather than the sheet, so on a short
transcript it floated in empty space at the window's corner. It now rides the
sheet, and drops clear of the bar when the layout is flipped.
Bottom clearance read `--composer-measured-height`, a surface var that never
reaches here, so it silently fell back to the root estimate and reserved ~20px
more than the bar occupies. The bar's real height is measured alongside the
band's.
Top-edge mode was never re-checked after the sheet started sizing itself: the
sheet stayed pinned to the bottom while the transcript hung from the top. Both
the anchor and the measurement now flip with the layout.
The model pill shed its label at 440px and the row stacked at almost the same
width, so the chevron bought nothing — which is the one thing a progressive
collapse is meant to avoid.
Sized off what the controls actually cost rather than a guess. With the full
pill they take ~284px, so at 440 the inline input was ~156px against a 128px
minimum: a few words wrapped, and wrapping is what stacks the row. Compacting
at 560 sheds the label while the input still has ~276px, and spends the ~110px
the chevron frees on keeping the row single for another stretch.
Global, not HUD-only.
An empty session painted a full-height slab of glass with nothing in it. The
sheet now grows up from the bar to fit the transcript and caps at the window, so
a fresh chat is just the bar.
Measured from the topmost row down to the bar and written straight to the
element — it changes on every stream flush, and the sheet resizing must not
re-render the tree.
The renderer half. HUD mode reuses the app's own chat surface and only changes
the frame around it: no titlebar, no statusbar, no pane tree. The band and the
bar tile the window between them with no dead margins, and the band runs the
window's full height with the opaque bar sitting on its bottom edge, so there
is no seam to compute and none to drift as the composer grows.
Visibility is a WoW chat frame: the transcript shows while a turn is recent or
the composer has focus, then holds and fades. A bottom-anchored gradient mask
carries the fade on both the sheet and the text, and flattens away on focus so
nothing is dimmed while you are reading. Only the composer never fades — it is
the interface.
The window is mouse-transparent everywhere it isn't really there, so clicks over
the faded band reach the app behind it; `pointer-events` can't do this, since
the click never reaches the page at all.
A transparent, frameless, always-on-top window that renders the real chat
surface, so its composer is the app's composer rather than a lookalike that
drifts. Main owns the window, its remembered geometry, and click-through.
Leaving is a handoff, not a window close. The gateway binds a session's event
stream to exactly one socket, so a turn started in the HUD streams only there
and the app window hears nothing — no deltas, no turn-complete, no draft clear,
and nothing to poll for mid-turn. So the app re-resumes the session the HUD
ended on, which rebinds the transport, reconciles the transcript, and picks up
an in-flight turn. Main carries the session id across, since it is the only
party that outlives the HUD's renderer.
Drafts are per-renderer state backed by shared localStorage, and the map is
read once at module load, so two windows on the same session diverge the
moment either types. Adds the two verbs a handoff needs: `reloadPersistedDrafts`
to merge another window's writes in (keeping local attachments, which are never
persisted), and a draft-sync bus so a composer can be told to flush its live
text down to the stash or repaint from it.
Dispatched synchronously, unlike the focus bus — a flush has to complete before
the window that will read it is created.
Resuming a session that is mid-turn somewhere else collapsed the thread down
to the in-flight prompt, and the user's own message never appeared at all
until a reload.
Two wrong premises in the resume path, neither specific to any one surface:
`omit_messages` was read as "empty transcript" rather than "no transcript in
this response". Desktop asks session.resume/activate to omit messages because
REST is the transcript authority, so mid-turn the live projection got
reconciled against an empty list and rebuilt the thread out of itself. The
response already carries `messages_omitted`; nothing read it. It now grafts
the projection onto the cache (or the REST prefetch) instead.
The settle path skipped hydration whenever the window had streamed the reply,
on the assumption that streaming a reply means owning the whole turn. True for
a turn you started, false for one you adopted: it arrives reply-first with no
prompt row, and nothing ever backfilled it. Sessions now carry
`adoptedRunningTurn`, set when a resume lands on an already-running turn and
consumed when it settles.
The windows-compat change-detector checked for the literal string
'start_new_session=True', but the systemd scope isolation path
conditionally uses start_new_session=False (the scope creates its own
session/cgroup). Assert 'start_new_session=' instead — the value may
now be a variable.
The entry_data/lock_held path (used by mark_turn_active/clear_turn_active)
was eagerly constructing a full O(n) snapshot of all routing entries under
the lock on every turn, even when the DB upsert would succeed. Defer the
fallback_data construction to the except branch where it's actually needed,
and respect lock_held to avoid re-acquiring a non-reentrant _lock.
When Hermes runs as a systemd gateway with MemoryHigh/MemoryMax limits,
local background terminal commands (terminal(background=true)) inherit the
gateway's cgroup. A memory-heavy executor (Codex, tests, Node) can push
the whole cgroup past MemoryMax and trigger systemd-oomd to kill the
ENTIRE gateway — taking down the messaging control plane and silently
losing the active turn.
Root cause: tools/process_registry.py::spawn_local() uses
start_new_session=True (creates a process session/group, NOT a resource
cgroup). The spawned process tree stays in the gateway's systemd cgroup.
Fix: when running under a service manager (detected via the existing
is_gateway_supervisor_process() helper), wrap the pipe-mode spawn command
in 'systemd-run --user --scope --unit=hermes-worker-<id>' so the worker
gets its own transient cgroup. An OOM in the worker then kills only the
worker, not the gateway.
The systemd-run availability is probed once (a no-op /bin/true in a
transient scope) and cached, because the binary can exist on PATH while
the user D-Bus session is unavailable (system services, containers). If
unavailable, fall back to the current start_new_session=True behavior
with a debug log.
Scope: this covers the common background pipe-mode path. PTY mode
(PtyProcess.spawn) is left as future work — it uses a different spawn
mechanism and is used for interactive CLI tools where cgroup isolation
has additional considerations.
- Fix _usage_audit_path() to use _get_hermes_home() instead of hardcoded
Path.home() / '.hermes' (profile-safe resolution, sweeper finding)
- Rewrite skip_background_review tests to exercise finalize_turn() directly
instead of duplicating the guard expression (sweeper finding)
- Fix response_silent audit field to use _is_cron_silence_response()
instead of the buggy SILENT_MARKER substring check it was meant to
replace (simplify-code review finding)
- Remove dead 'model' in locals() guard — model is always in scope
before the try block (simplify-code review finding)
- Extract _stub_agent_for_finalize() helper to eliminate ~40 lines of
copy-pasted agent stubbing in tests (simplify-code review finding)
- Clean up 'Phase 0.5' instrumentation comments
Phase 0.5 of the Hermes Agent token leak mitigation plan: append a single
JSONL line to ~/.hermes/cron/usage_audit.jsonl after every cron LLM
invocation, capturing prompt/completion/total tokens, model, duration_ms,
deliver target, and error (when raised). Read from agent.session_*_tokens
which run_conversation already returns in its result dict.
Without this, we have no measured baseline to attribute token deltas to
subsequent mitigation phases. The plan's hard gate: observability lands
before any mitigation phase.
Writer NEVER raises — wrapped in a single try/except that logs a warning
on any json.dumps / mkdir / open failure so an audit-log bug cannot
break a cron job. Failure-path audit guard via locals() check covers
exceptions that fire before the fire_id is assigned.
No new dependencies, no new env vars (the plan rejected one in v2).
Tests: 7 new unit tests in tests/cron/test_usage_audit_logger.py covering
the success path, missing token info, swallowed writer exception, parent
dir creation, multiple appends, and unicode preservation.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Phase 8 wire-in + Vector 8 doc comment from
ralplan-hermes-token-leaks.md.
(1) Phase 8 wire-in: cron AIAgent construction now passes
skip_background_review=True. This suppresses the end-of-turn
skill/memory review fork (~30K tokens/event, ≤30K typical and
≤150K worst-case daily on bluenode) which has no human-in-the-loop
value for cron sessions.
(2) Vector 8 doc comment: a one-line comment immediately above the
AIAgent(...) construction documenting the verified-negative
finding that title generation does not run on the cron path
(maybe_auto_title is gateway/CLI-side only). Future contributors
won't accidentally introduce title-gen here without realizing it
would add ~600-1000 tokens/fire on a path that explicitly opts out
of memory/review/title overhead.
No new tests required for the doc comment (no behavior change). The
skip_background_review wiring is covered by the existing source-text
assertion in tests/agent/test_skip_background_review.py.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Phase 8 of the Hermes Agent token leak mitigation plan
(ralplan-hermes-token-leaks.md §3.9). Adds a boolean kwarg
`skip_background_review` (default False) to AIAgent.__init__ that
suppresses the end-of-turn _spawn_background_review fork.
Each background review fork instantiates a new AIAgent with its own
~15K input tokens + up to 8 LLM iterations, accumulating ~30K tokens
per event in the worst case. On cron sessions there is no
human-in-the-loop benefit from the review (no skill-creation pressure,
nobody curating MEMORY.md), so the cost is pure waste.
The end-of-turn guard now reads:
if (final_response and not interrupted
and not getattr(self, "skip_background_review", False)
and (_should_review_memory or _should_review_skills)):
skip_memory=True already disables the memory-review trigger; this
flag is the explicit single-switch off for both review paths.
Defaults to False, so behavior is unchanged for gateway/CLI callers
that omit the kwarg.
Tests: 5 new unit tests in tests/agent/test_skip_background_review.py
covering the default value, flag persistence, the gate short-circuit,
the gate fall-through, and a source-text assertion that the cron
scheduler sets the flag to True (separate commit).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Table-driven across every session-scoped RPC that can hit a dead runtime
id, plus the invariants the scattered copies disagreed on: resume targets
the session-owning profile (never forks into the active one), the recovered
id publishes exactly once, drift during the resume aborts instead of
retrying, a resume that itself 404s rethrows the ORIGINAL error, and
recovery is bounded to a single retry.
These run without touching the REST layer because profile resolution is
injected — the pre-consolidation helper reached through resolveStoredSession
-> getSession(), which made its coverage depend on store state left behind
by whichever test file ran first (passed alone, timed out at 15s alongside
index.test.tsx).
Co-authored-by: xxxigm <xxxigm@users.noreply.github.com>
Co-authored-by: rapsealk <rapsealk@users.noreply.github.com>
Co-authored-by: webtecnica <webtecnica@users.noreply.github.com>
Same bug class, remaining call paths. runRewindSubmit handled only
"session busy" — a rewind runs right after cancelRun, and interrupting can
drop the gateway's session, so "Restore checkpoint" after a stop hit a dead
runtime id and surfaced the raw error. The session-tile delegate's
interrupt/submit had no recovery either, so a tile left open across sleep
was dead until reopened.
Both now use the shared resolver. The tile delegate resolves its durable id
by reversing the stored->runtime cache and repoints that mapping on
recovery, so later tile actions use the live binding instead of recovering
on every call.
Co-authored-by: JonthanaHanh <JonthanaHanh@users.noreply.github.com>
Co-authored-by: akivavh <akivavh@users.noreply.github.com>
After sleep/wake, a long idle, or a remote backend restart, the gateway
drops its in-memory runtime session while the desktop still holds the old
id. prompt.submit already recovered, so plain text kept working and the
failure looked selective: attaching an image or running /compress died
with a bare "session not found" and a new chat was the only workaround.
Attach runs BEFORE prompt.submit, so submit's recovery never got a
chance — image.attach_bytes / image.attach / file.attach failed first.
Route them (and session.compress) through the shared resolver, and thread
the recovered id back to the caller so the follow-up submit targets the
live session instead of the dead one.
Attachment bytes are read once, outside the retry, so recovering a large
upload doesn't re-read it. /compress deliberately does NOT opt into
timeout recovery: it is legitimately LLM-slow and retrying a timeout
would double a minutes-long call.
Reported-by: bapemonkey (Discord)
Co-authored-by: xxxigm <xxxigm@users.noreply.github.com>
Co-authored-by: zzz163519 <zzz163519@users.noreply.github.com>
Co-authored-by: luxles <luxles@users.noreply.github.com>
main carried three hand-rolled copies of the same recovery policy —
prompt.submit (submit.ts), session.interrupt (cancelRun) and
session.redirect (steering) each open-coded
isSessionNotFoundError -> resolveSessionProfile -> session.resume ->
retry. Three copies of one policy is how call sites drift apart, and it
is why the RPCs that were never given a copy (attach, /compress,
checkpoint restore) still surface a raw "session not found" after
sleep/wake while plain text silently recovers.
Introduce withSessionNotFoundResume() as the single resolver and move
all three existing copies onto it. Profile resolution is injected rather
than imported so the helper is unit-testable without reaching through
resolveStoredSession -> getSession(); a drift callback lets each caller
keep its own abort semantics via SessionRecoveryAborted.
Co-authored-by: xxxigm <xxxigm@users.noreply.github.com>
Eight cases across the whole class, not just the reported path:
- cold resume rebinds from the selected row before resume settles
- an empty runtime cwd releases ownership (the permanent-staleness half)
- releasing leaves the PATH intact, so panes don't collapse and the persisted
workspace survives
- a session row outside the loaded sidebar page doesn't blank the pane
- a non-git workspace with a null git_repo_root still uses its row cwd
- the branch label clears so the previous project doesn't leak
Verified as a real barrier: reverting utils.ts fails 5 of these.
Co-authored-by: xxxigm <tuancanhnguyen706@gmail.com>
Co-authored-by: ZHJay <ZHJay@users.noreply.github.com>
Co-authored-by: worlldz <worlldz@users.noreply.github.com>
`_fallback_session_info` returned `_default_session_cwd()` — the directory the
gateway process happened to start in — so a session resumed without a built
agent told its client the wrong workspace, and the desktop Files pane painted
the wrong project even after the renderer rebound correctly.
Return the session's own cwd and always emit `branch` ("" outside a git repo)
so a client can clear a stale label instead of retaining it. This matches the
contract `_lazy_session_info` already follows a few hundred lines above.
Co-authored-by: ZHJay <ZHJay@users.noreply.github.com>
`session.info` claimed the cwd for whatever the selection happened to be, so a
background tile's payload could re-point a fresh draft at the tile's workspace.
Treat a nonempty `stored_session_id` as non-matching when no primary session is
selected; only an ABSENT id uses the selected-session fallback (the backend
omits it on a lazy session, and refusing there would leave the workspace
un-owned for the rest of the conversation).
Matching goes through the lineage rather than raw string equality: the backend
id is the live session_key, which auto-compression rotates to the continuation
tip, while a selection made from a pinned row holds the stable lineage root.
Comparing those literally reads one conversation as two.
Co-authored-by: ZHJay <ZHJay@users.noreply.github.com>
Two defects left the Files pane showing the previous project's tree:
- `applyStoredSessionPreviewRuntimeInfo` reset every composer atom EXCEPT cwd,
and runs before the `session.resume` RPC. The sidebar row already knows the
conversation's workspace (`cwd` is in the compact row projection), so mirror
it on the same tick the selection changes.
- `if (info.cwd)` was truthy-only, so a detached session reporting `cwd: ''`
never cleared and the pane stayed pinned to the last project for the rest of
the session — the "not always" in the report. Empty is now authoritative.
Empty routes through ownership release rather than a persisted `''`:
`setCurrentCwd` writes to localStorage and seeds `$currentCwd` on next boot, so
blanking would also wipe the remembered workspace.
Only `cwd` is consulted, never `git_repo_root` — the latter is documented null
for non-git workspaces and not-yet-backfilled rows, so falling back to it reads
as "no workspace" and blanks a pane that was correct. A session outside the
loaded sidebar page (no row at all) releases ownership instead of blanking, for
the same reason.
Also claims ownership on the warm-cache path (its missing-RPC compat branch
returns before `applyRuntimeInfo`) and for a center tile, whose Project "+"
create left the right rail on the previous session's folder.
Co-authored-by: xxxigm <tuancanhnguyen706@gmail.com>
Co-authored-by: worlldz <worlldz@users.noreply.github.com>
Co-authored-by: ZHJay <ZHJay@users.noreply.github.com>
`$currentCwd` is a global singleton, but a conversation switch publishes the
new stored session id immediately while the new workspace only arrives when
`session.resume` settles. For that whole window the path still names the
PREVIOUS conversation, and every workspace-derived surface treats it as
authoritative.
Track WHICH conversation the live path describes instead of trying to keep the
path itself in lockstep. Ownership — not emptiness — is what makes the switch
atomic: clearing the path would collapse the workspace/review panes and drop
file-tree state on every switch, so the path stays put and is simply marked
not-yet-owned.
The released marker is deliberately not `null`: `null` MATCHES a fresh draft
(whose selected id is also null), so releasing to it would hand a leftover path
to the draft as its own workspace.
Co-authored-by: ZHJay <ZHJay@users.noreply.github.com>
The main-checkout test compared the two probe roots with raw string
equality. When they differed only in separator spelling, the repo's own
checkout was misclassified as a linked worktree: it fell through to the
worktree branch and was labeled by directory basename. The sidebar then
showed one checkout twice — a dir-labeled lane plus the branch-labeled
`main` lane built from the same sessions.
Compare with `_path_key` so platform path identity decides, matching how
every other path comparison in this module is already keyed.
Tests cover the single-checkout case and the main + linked-worktree case;
both fail before this change (the lane comes back labeled `repo`, not
`main`).
`common_repo_root` derives its answer via `os.path.realpath` +
`os.path.dirname`, which rewrite separators to the platform-native `\` on
Windows, while `repo_root` returns raw `--show-toplevel` output (always
forward slashes). The same directory therefore came back spelled two ways
from a single `resolve()` call, so callers comparing the two roots for
identity could not see that a repo's own checkout IS its common root.
Normalize the derived path back to git's forward-slash spelling so both
probes agree byte-for-byte.
Fleet audit showed these task skills are commonly needed across users;
shipping bundled per Teknium's direction. Docs and tests follow the
bundled paths.
Brooklyn Nicholson