A SessionDB handle cannot be released by dropping the last reference.
Once its background token writer starts, the instance pins ITSELF two
ways: the writer thread's target is a bound method, and
queue_token_counts registers atexit.register(_drain_token_queue_at_exit),
which only close() unregisters. A dropped-but-pinned handle keeps its
state.db/-wal/-shm descriptors for the life of the process, and __del__
never runs for it, so the existing safety net is dead code for exactly
the instances that leak.
That is why owning call sites are expected to close explicitly, in those
words, in the ownership comments in run_agent.py and
tui_gateway/methods_session.py. This adds the ergonomic half of that
contract so an owner can scope a handle and be exception-safe by
construction:
with SessionDB(path) as db:
db.append_message(...)
Purely additive. __enter__ returns self, __exit__ closes and returns
False so a caller's exception always propagates, and close() is already
idempotent, so a scope that closes early still exits cleanly. Nothing
changes for callers that already close directly.
Four regressions cover the scope closing the handle, __enter__ returning
the instance itself, the failure path closing while still propagating,
and an early close leaving the exit clean. They assert on the
sqlite_safe_read tracking registry rather than raw descriptor counts,
matching test_session_db_read_conn_pool.py, because SQLite's unix VFS
parks a closed descriptor on a per-inode reuse list and makes raw counts
lag the real connection count.
Refs #88033
Since the managed-cron redesign (#84339, v2026.8.13) the dashboard fire
webhook forwards fires to the gateway process and returns 503 when it is
unreachable so NAS/QStash retries. Correct for transient windows — but an
operator-STOPPED gateway can never be fixed by retrying: every fire on
every job burns the full scheduler retry budget, NAS converts each 503 to
a retryable 502, and the resulting storms page on-call for a non-incident
(OOF-266 and its five duplicate tickets; +93% relay callback failures as
the fleet adopted v2026.8.13).
Split the unreachable path by durable operator intent:
- desired_state == "stopped" (written only by the s6 lifecycle commands;
the same intent signal container-boot reconciliation trusts) -> drop
the fire with 200 + a structured log line, mirroring NAS's own
instance_stopped drop. Jobs are not lost: the Chronos provider
reconciles and re-arms every job on the next gateway start.
- Anything else (crash loop, scale-to-zero wake, restart, legacy state
file without desired_state) -> keep the retryable 503, now stamped
with Retry-After: 60 so a scheduler that honors it spaces retries
past the wake/restart window instead of exhausting them inside it.
The gateway's own pass-through 503s (draining) get the same hint.
The intent check fails open (any parse/resolution error -> retryable
path) and is only consulted when the gateway is actually unreachable, so
a stale state file can never shadow a live gateway.
OpenAI enabled the large-context window for ChatGPT-subscription Codex
accounts (announced by @thsottiaux Aug 16 2026; previously API-key-only).
Live re-probe the same day: 911,276 input tokens completed OK on
gpt-5.6-sol; ~925K+ rejected with context_length_exceeded (1.05M window
minus reserved output headroom). terra, luna, and gpt-5.4 all completed
900,026 tokens OK. The Codex catalog still advertises 272K, so the
stale-advertisement override from #87981 is the right lever — this just
raises its value 350K -> 900K.
gpt-5.5 and gpt-5.4-mini still enforce 272K live (rejected 500K) and
remain excluded. Override semantics unchanged: fires only on an
exactly-272,000 advertisement; any live catalog change is trusted
verbatim.
Pulls the multi-source roster into the bundled plugin: profiles.list rows
from the active gateway are merged with the host.agents() union roster
(hermes-agent #86875), so the Bots panel shows agents from every registered
Desktop connection with @name-device handles for duplicates. Feature-detected
and best-effort — an older Desktop build or roster failure leaves the
single-source list untouched.
Adapted for the bundle:
- useRoster queryFn combines the bot_mode_protocol capability read (which
landed after #68 was cut) with the multi-source merge
- multi-source-roster tests updated for the namespace SDK import harness
- soul-protocol-backfill anchor widened for the new botHandle(name, bot)
signature
Plugin suite: 143/143.
The config-read guard (test_config_read_guard) correctly flagged the
probe's raw yaml.safe_load of config.yaml — raw reads miss the managed
overlay, env expansion, and normalization. Use load_config_readonly()
under a scoped HERMES_HOME override instead. E2E v3/v3b and the guard
both green.
Bot Chats created before the epoch mechanism persisted prompts with no
protocol section and no stamp — the staleness check only fires on
stamped prompts, so pre-existing bots would never learn to message
teammates. stored_bot_chat_prompt_needs_upgrade() migrates them: one
rebuild, title-gated to Bot Chat, only when the probe would actually
emit a section (SOUL-append legacies and unmanaged installs are left
alone — rebuilding those would loop). The rebuilt prompt carries the
stamp, so the upgrade can never re-fire.
E2E v3b through the real restore path: legacy Bot Chat upgraded once
then verbatim-reused; legacy regular sessions byte-untouched.
tests/agent/ 4648/4648.
Bot Chats break the "new sessions come often" assumption behind
build-once system prompts: capability edits used to sit invisible until
/new or compression, and the frozen birth date became misinformation.
- tools/bot_mode_probe.py: capability_fingerprint() hashes the profile's
capability surface (disabled skills, toolset pins, MCP config, SOUL.md,
installed skills, Bot-Mode roster); Bot Chat prompts embed the 12-hex
epoch stamp
- agent/conversation_loop.py restore path: stored Bot Chat prompt whose
epoch mismatches disk → ONE rebuild (through a cleared skills-prompt
cache so new installs appear), persisted so the next turn reuses the
new bytes verbatim. Prompts without a stamp — every non-Bot-Chat
session — never take the branch; probe failure fails closed to reuse
- agent/system_prompt.py: Bot Chat prompts are timeless — the
"Conversation started:" date is dropped (timezone kept); no ticking
fields in an eternal session
- tui_gateway: _sync_bot_capabilities at turn start rebuilds the live
agent (tool definitions are construction-baked) when the fingerprint
moves, same session id/history, with a user-visible notice
Cache stance: this is the /model exception applied to capabilities — a
loud, user-initiated, once-per-change prefix break. Unchanged state
hashes identically and stored bytes are reused verbatim (E2E-proven).
Validation: 9 probe unit tests incl. per-axis fingerprint changes;
E2E v3 against the real restore path (fresh build → verbatim reuse →
skill install → single refresh w/ new skill in index → verbatim reuse;
regular sessions dated, unstamped, never refreshed); tests/agent/
4647/4647.
Live desktop E2E caught a write-ordering bug the automated E2E missed:
tui_gateway applies pending_title to state.db AFTER the first turn, but
the system prompt builds at turn START — the DB-title gate saw nothing
and the Bot Chat was cached protocol-less forever. The gateway now
hands the agent its intended title at construction and the gate checks
the hint first, DB second (CLI/messaging-gateway paths unchanged).
Live-verified on the running desktop: fresh bot's Bot Chat persisted
with the protocol section, handle, and roster in its system prompt;
regular sessions and SOUL.md untouched.
Found in live desktop E2E: the generated-identity path of composeSoul
still appended the protocol section even when the backend injects it
into the system prompt. New agents now get a clean identity-only SOUL
against capable backends; older gateways keep the append. Covered in
the capability-suppression test.
Per review: the protocol belongs only in official Bot Mode interactions,
not every session on a managed install. The prompt builder now injects
the section only when the agent's session row is titled "Bot Chat"
(BOT_CHAT_TITLE, matching the desktop's createCanonicalChat pin and the
`hermes -p <bot> chat -c "Bot Chat"` resume target). Regular sessions
never carry it; the desktop composer middleware owns @mention sends.
Title is read once at first prompt build and the rendered prompt is
cached + DB-restored — cache-safe. E2E against the real AIAgent +
SessionDB: absent in an untitled session, present in Bot Chat,
byte-stable across rebuilds, absent after retitle, absent with the
flag off. Overhead unchanged (~916B, Bot Chat sessions only).
apps/desktop/src/**/*.js is gitignored (stale tsc output shadows .tsx),
which silently dropped the hermes-bots plugin.js from the adoption
commit — tests shipped, source didn't, CI ENOENT'd. Negate the pattern
for src/plugins/*/plugin.js: adopted plain-ESM plugins have no .tsx
sibling, so the shadow hazard cannot apply.
Replaces the plugin-side SOUL.md protocol append: on Bot-Mode-managed
installs (any profile carrying ui_meta['hermes-bots']) the prompt builder
injects the "Messaging other agents" section into every session of every
profile — including headless `hermes -p <bot> chat` sessions a teammate
starts — so bot handoffs work without mutating user-authored SOUL files.
- tools/bot_mode_probe.py: silent-when-unmanaged probe, cached per
(process, home), keyed off the agent's OWN home (not ambient
HERMES_HOME); silent when SOUL.md already carries the legacy section
- agent/system_prompt.py + agent_init.py + config_defaults.py: wired as
agent.bot_mode_protocol (default True), stable tier, byte-stable
across rebuilds (E2E-verified against the real build_system_prompt)
- tui_gateway profiles.list gains bot_mode_protocol capability flag;
the bundled plugin gates ALL SOUL protocol writes on it (backfill,
composeSoul, Edit save) — older gateways keep the SOUL-append path
- overhead: ~916 bytes, only on Bot-Mode installs; zero elsewhere
Supersedes the SOUL backfill half of Hermes-Bot-Mode#99 (credit
@kaduxo — the handle fix, `hermes profile list` correction, and
idempotent-append guards from that PR ship in the bundled plugin).
Adopts the Hermes-Bot-Mode desktop plugin (NousResearch/Hermes-Bot-Mode)
into apps/desktop/src/plugins/hermes-bots/, registered by the bundled
vite glob and ON by default. It stays a pure @hermes/plugin-sdk consumer
in plain-ESM plugin.js form; users disable it live in Settings > Plugins.
- contrib/plugins.ts: bundled glob accepts plugin.js entries
- contrib/runtime-loader.ts: a disk/runtime copy of an id that ships
bundled is skipped (standalone installs predating adoption cannot
double-register)
- package.json: check:test:plugins runs the plugin's node:test suite in
CI (138 tests)
- source: Hermes-Bot-Mode @ c19baba, incl. today's #107/#103/#99 merges
The live suite previously called appendMidTurnUserMessage directly, leaving
redirectPrompt's appendAfterActiveReply guard — the production decision of
WHERE a correction lands — outside the harness. Both hooks now mount together
sharing one state map, exactly as the desktop wires them, so a regression in
the caller (not just the insert) goes red. Verified by mutation: disabling the
guard fails 2/4.
Also covers the rejected-redirect path: a not_running response discards the
optimistic bubble instead of stranding a correction the model never saw.
The hydration fixture now carries the durable row shape the client actually
receives (row_id, reasoning, provider call_id/response_item_id on tool_calls)
instead of a hand-simplified echo, so the 'mirrors real state.db rows' claim
is honest. The fake-timer steer id counter is gone with the local insert —
ids come from redirectPrompt itself.
Review follow-ups: steer ids now come from a monotonic counter instead of
Date.now() (frozen under fake timers — two steers without a clock advance
would have collided), and the settle-above assertion documents its
load-bearing sealed-bubble assumption.
A steered turn's contract — pre-steer output above the correction bubble,
post-steer output and the settled reply below it — was fixed across
several PRs (#73793/#83151 class, settle fixes) but only covered piecewise:
the mid-turn insert as a unit, the settle math as a unit. Nothing drove the
real stream reducer through a whole steered turn, and nothing asserted the
durable-row hydration renders the same order after reload.
Two suites close that:
- steer-arrival-order: full event sequences through useMessageStream's real
handler + the real optimistic insert — single steer with tool activity,
steer racing message.complete, double steer in one turn.
- steered-turn-hydration-order: toChatMessages over persisted row shapes
copied from a real state.db steered turn, including a tool result that
lands after the correction row.
When Hermes Desktop works against a REGISTERED gateway connection, cron
jobs execute on that gateway and persist their run sessions in the
gateway's state.db. But every REST call in the app — the cron surface
included — carried only `profile`, so `hermes:api` routed it through the
local profile pool and `_list_cron_job_runs_sync` read a local state.db
with zero `source='cron'` rows. Every job showed "No runs yet" while the
same endpoint on the gateway returned the real runs (#87882).
Fix at the routing seam:
- HermesApiRequest gains an optional `connectionId`. The renderer's cron
helpers (list/get/runs/delivery-targets/create/update/pause/resume/
trigger/delete/blueprints) now tag the active registry connection via a
new connectionScoped() twin of profileScoped(), fed from the same
setApiRequestConnection seam store/gateway already maintains for the
plugin socket.
- The hermes:api main-process handler resolves a tagged request through
ensureRegistryBackend — the SAME pool the job list and WS traffic use —
instead of the legacy profile route. Shared remote/cloud hosts (one
gateway, many profiles) get the path scoped with ?profile= via the new
pathWithProfileScope helper, factored out of pathWithGlobalRemoteProfile.
- '' / 'local' / absent connectionId keep the byte-identical v1 route, so
single-source and connection-config-remote users are unaffected.
This covers the run-history panel, the sidebar cron peek, and every other
cron surface in one place, since they all funnel through the same helpers.
Fixes#87882
A connection/mode apply (soft re-home) moves /api/profiles routing to a new
backend, but nothing deterministically re-fetched the rail's $profiles list
and a stale in-flight response from the previous backend could land last and
collapse the rail to Home (#85731).
- store/profile: epoch-guard refreshProfiles/refreshActiveProfile so a
response fetched against the previous backend never writes the shared cache
(invalidateProfileListFetches), and bump the epoch on live profile swaps.
- store/gateway-switch: strand in-flight profile-list fetches in the same
wipe every connection/mode apply funnels through.
- use-gateway-boot: explicitly re-pull the active profile + list from the NEW
backend during softSwitch, best-effort like its sibling fetches.
Fixes#85731
Multiple Desktop windows share one renderer origin (one localStorage
area) while each window can be connected to a DIFFERENT gateway. The
sidebar pin set (hermes.desktop.pinnedSessions), the manual session
order, and the remembered last-session/route navigation keys were all
persisted under single global (or profile-only) keys, so two windows on
different gateways read and reconciled the same lists: pin-sync's
pullRemotePins() in one window adopted/dropped pins belonging to the
other window's backend, producing the overlapping mixed PINNED/SESSIONS
lists reported after the v0.19.1 update relaunch.
Introduce a connection-scope persistence layer (connectionScopedAtom in
src/lib/connection-scoped.ts): the local connection keeps the bare
legacy key (byte-identical for single-backend users, same contract as
backendScopeKey), while remote connections persist under
`<key>.remote.<encoded baseUrl>.<encoded profile>` — the shape
workspaceCwdKey already established. setConnection() rescopes every
scoped atom when the window's connection changes (null descriptors keep
the current scope, as with syncCronModelImpactConnection), and pin-sync
resets its mirrored/pending/unconfirmed bookkeeping on rescope so a
reconcile never PATCHes one gateway's pins to another.
Legacy globally-keyed values are deliberately not migrated into remote
scopes: ownership of rows accumulated by every window is unknowable
(the #67709 precedent), and backend-mirrored pins self-heal from the
gateway's own `pinned` rows.
Fixes#77318
* fix(tui): restore Alt+Enter for newlines
Restore Alt+Enter support for inserting a new line in the TUI after the behavior was lost during newer input-handling updates.
Legacy terminals encode Alt+Enter as ESC followed by carriage return. Preserve those bytes as a single tokenizer sequence and parse the result as Return with the Meta modifier so TextInput inserts a newline instead of submitting.
Keep plain CR and LF mapped to unmodified Return, and cover the legacy ESC+CR sequence with a regression test.
* fix(tui): scope legacy Alt+Enter tokenization
The Codex /models catalog advertises 272K for the gpt-5.6 (sol/terra/luna)
and gpt-5.4 slugs, but the backend actually accepts ~371K input tokens
(verified live against chatgpt.com/backend-api/codex/responses, Aug 16 2026:
~371K completed OK on all four slugs; ~382K+ rejected with
context_length_exceeded). 350K keeps ~22K margin under the observed ~372K
enforcement.
The bump applies ONLY when the resolved value is exactly the known-stale
272,000 advertisement — any other advertised value (higher or lower) is
trusted as a real server-side change, so a future catalog correction
deactivates the override automatically. gpt-5.5 and gpt-5.4-mini both
genuinely enforce 272K (rejected 360K live) and are excluded.
The runtime-contract repair now also runs during hermes update and once
per session at the first computer_use call (PR #87923); the docs only
mentioned setup and toolset enablement.
A same-day version-floor bump (0.20 runtime contract) left every install
with an older cua-driver hard-failing on all computer_use calls: the
start() gate fails closed, while the `hermes update` refresh defers to the
driver's own check-update verb — whose ~20h cache routinely answers "no
update available" right after we raise the floor. Hermes knew it required
0.20+ but never acted on that knowledge.
Two changes:
- tools_config.install_cua_driver(): a contract-failed installed driver is
repaired on the upgrade=True path too (previously only upgrade=False).
The contract failure itself is the confirmation, so the
require_confirmed_update gate and the check-update short-circuit are
bypassed for repairs — an indeterminate or stale-cached check can no
longer pin users on an unusable driver.
- cua_backend.CuaDriverBackend.start(): when the contract gate fails on an
installed binary, attempt one automatic repair per process via the
standard install path, then re-probe. HERMES_CUA_DRIVER_CMD overrides
are never repaired (explicit override is authoritative even when broken)
and a missing binary still just reports the install hint. A failing
installer can't loop: the second start() surfaces the original error.
Tests: contract-repair coverage in test_computer_use.py (auto-repair
success, failed repair surfaces the original error, once-per-process
guard, override never repaired, missing binary never repaired) and
test_install_cua_driver.py (incompatible driver repairs despite an
indeterminate check-update, check-update not consulted). All new tests
verified to fail against the unfixed source (sabotage run).
With a GH_TOKEN/GITHUB_TOKEN in the environment, electron-builder auto-selects
the github provider and resolves owner/repo from the repository field, falling
back to reading <projectDir>/.git/config. projectDir is apps/desktop, which has
no .git of its own, and app-builder-lib does not walk up to the workspace root
-- so resolution returned null and threw "Cannot detect repository by
.git/config".
On Linux this fires from onAfterPack for a plain `dir` target: the darwin and
Windows branches return early for non-installer targets, Linux has no such
guard. That is why the same build worked elsewhere.
--publish never keeps `pack` from reaching this at all, but `dist:*` and
test-desktop.mjs still resolve publish config on a machine with a token, so
declare the field too.
Tests call the real app-builder-lib resolver rather than asserting on the text
of package.json, so they track electron-builder's behavior instead of our
formatting.
Co-authored-by: airo7 <airo7@users.noreply.github.com>
Co-authored-by: frankmendes1979 <frankmendes1979@users.noreply.github.com>
`hermes desktop` runs `npm run pack` through _npm_lifecycle_env(), which
sets CI=1. electron-builder 26 reads that as an implicit publish request
(`onTagOrDraft`) when --publish is absent, so a local --dir build enters
publish resolution it has no business being in.
Pin `--publish never` on the pack script. This is also what electron-builder
asks for directly -- the implicit CI behavior is removed in v27.
Co-authored-by: webtecnica <webtecnica@users.noreply.github.com>
Co-authored-by: fangliquanflq <fangliquanflq@users.noreply.github.com>
`--yolo` / `-z` read as "don't prompt me", but they also swap computer_use
onto a private `unrestricted` daemon, dropping the ceilings the configured
mode would have applied. Nothing said so. A script picks up `-z` for quiet
output and loses its limits as a side effect, and the only trace is a driver
process nobody inspects.
The mapping itself stays. It is deliberate, and `unrestricted` is reachable
no other way: it is intentionally not a config value so a stale config line
can never silently bypass approvals (see `_cua_configured_permission_mode`).
Removing the mapping would delete the capability rather than fix it, and
splitting it onto a second CLI flag was declined to avoid growing the
surface.
So the widening is now stated instead: one warning per session naming the
configured mode it left, what stopped applying, and the two ways to keep a
ceiling - drop the bypass flag, or declare a version-3 capability manifest,
which now rides along with unrestricted as of the previous commit.
Once per session, not per dispatch: the resolver runs on every tool call.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
`--yolo` / `-z` route the session onto a private embedded daemon in
`unrestricted` mode. That daemon was constructed without the configured
capability manifest, and the serve command only attached
`--capability-manifest` when the mode was exactly `bounded`. So the moment a
run was bypassed, the user's declared ceiling was dropped:
without -z: --permission-mode bounded --capability-manifest ...
with -z: --permission-mode unrestricted --dangerously-bypass-approvals
No manifest, no warning. The most carefully configured run - a reviewed
ceiling, written by hand - became the least constrained one, silently, and
it failed open.
That was never a driver limitation. cua-driver documents the manifest as a
ceiling across modes ("A manifest can narrow a profile but never widen it";
its own authorization table calls it `optional_capability_manifest_ceiling`),
and accepts it alongside `--permission-mode unrestricted`.
The forwarding is version-aware, because the two manifest schemas differ
(cua-driver session_manifest.rs):
* v1/v2 are legacy and must declare `mode: bounded`. Handing one to an
unrestricted runtime aborts startup with "legacy capability manifest mode
must be bounded", so a naive forward would turn a working session into a
hard failure. These are forwarded for bounded only, and a warning names
the migration when one cannot apply.
* v3 must not declare a mode. It is the mode-independent ceiling, and it now
rides along with unrestricted.
Unreadable or unparseable manifests are not forwarded outside bounded, on
the same fail-safe reasoning; bounded still forwards unconditionally and
lets the driver be the authority there.
Verified against cua-driver 0.20.0 on Windows. Launch args now carry
`--permission-mode unrestricted --dangerously-bypass-approvals
--capability-manifest <v3> --approve-capability-manifest`, and the ceiling
is enforced in the bypassed run - a tool outside the manifest is refused
("outside the capability manifest for this session ... blocked as a
protected resource") where the same config previously ran unbounded. A
legacy manifest was confirmed to abort driver startup when forwarded, which
is what the version gate prevents.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
`cua_browser_state` has two branches, chosen implicitly: any call carrying
pid or window_id is a *binding* (browser_route.py:252), anything else is a
*snapshot*. A binding clears session state, mints fresh tab_ids, returns
binding metadata with no page content, and sets verification_required.
Nothing in the response says that. A caller that keeps passing pid/window_id
- the natural reading of "bind to this window, then read it" - re-binds
forever: the tab_id it just received is unbound by the next bind, so every
cua_browser_navigate comes back browser_verification_required, and the
refusal ("take a fresh snapshot") points at the same call that just re-bound.
Observed live as 11 consecutive refused navigates before the model gave up
and fell back to foreground SendInput on the address bar.
The same confusion silently swallowed include_screenshot: both calls that
requested one were bindings, which carry no page content, so the flag had
nothing to attach to and was dropped without comment.
A binding response now reports snapshot_required, next_step
(fresh_browser_state, matching the existing token convention) and a hint
naming the exact next call; requesting a screenshot on a binding reports
screenshot_deferred instead of dropping it. The verification refusal now
says to call cua_browser_state WITHOUT pid/window_id and why re-sending them
does not help. The schema documents that include_screenshot applies to
snapshots.
Behavior of the bind and snapshot branches themselves is unchanged - this is
purely about making the split legible to the caller.
Unit-tested. Not verified end to end on the reporting host: the driver
refuses the bind upstream there (`browser_requires_setup: no owned DevTools
endpoint`, and it does not accept a user-launched --remote-debugging-port),
so the typed route never reaches this branch. That attach failure is a
separate cua-driver issue.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Live-testing the Cua Driver 0.20 convergence on Windows 11 (session 2,
cua-driver 0.20.0) surfaced three defects in the existing-profile browser
path and in install status.
1. The config grant was silently nullified by an approval bypass.
`--yolo` / `-z` map onto a private unrestricted daemon, which answers every
browser_prepare. Because the host delegated the entire existing-profile
decision to the driver, that bypass also nullified
`computer_use.grant_existing_profile: false`: a plain `hermes -z` attached
to the user's real Chrome profile and read live page content over CDP, with
the driver reporting it as "the approved existing Chromium profile". It was
never approved.
An approval bypass is consent to skip prompts, not consent to read an
existing profile's pages, cookies, and storage. CuaTypedBrowserRoute.prepare
now enforces the key itself, regardless of permission mode. bounded stays
exempt - its reviewed capability manifest is the authorization boundary.
The authorization inputs are resolved in the backend from config and the
backend's immutable mode, never from model-supplied kwargs.
2. The grant, once set, still could not be used.
With `grant_existing_profile: true` the runtime is launched
`--grant existing-profile` correctly, but cua_browser_prepare then hit a
runtime approval prompt anyway - re-asking the user to authorize what the
config already authorized, and making the documented opt-in unusable on any
non-interactive run, where the prompt has nobody to answer it and the call
dies on approval timeout. The durable, file-backed grant now stands in for
that prompt. Scope is narrow: only the existing-profile prepare, only when
the grant is present; isolated launches still prompt and any resolution
failure falls closed to prompting.
3. `computer-use status` hid a custom override and spliced its output.
With HERMES_CUA_DRIVER_CMD pointed at cmd.exe, status printed the child's
multi-line banner and prompt inside the one-line version field, never
mentioned the override, and advised `hermes computer-use install` - which
install itself (correctly) refuses to run against an overridden path. It now
names the override and mirrors install's update-or-unset guidance, and
version output is reduced to one bounded line.
Verified on the reported host: `-z` existing-profile attach now refuses and
names the key; `grant: true` no longer prompts (33s vs a 300s approval
timeout); status names the override and prints one line. No change to the
reconciliation path - driver SHA256 unchanged end to end.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The sequential tool path only noticed a user interrupt after the running
tool returned: with the deadline disabled it ran the tool inline (fully
blocking), and with a deadline it waited in 5s slices without ever
checking agent._interrupt_requested. Any tool without cooperative
is_interrupted() polling (image_generate, tts, transcription, skills
sync, ...) held the whole turn hostage — the reported symptom was a
redirect queued ~40s behind a FAL image generation + upscale pass.
Executor backstop (class fix, covers ALL tools):
- _run_sequential_tool_execution_middleware always dispatches on the
daemon worker (timeout None no longer means inline blocking) and polls
the interrupt flag every 1s.
- On interrupt: 3s cooperative grace (mirrors the concurrent path), then
synthesize a cancelled tool result (_ToolCancelledResult), emit the
terminal post_tool_call with status=cancelled, and abandon the worker.
- _ToolCancelledResult suppresses downstream post-hook double emission
exactly like _ToolTimeoutResult, so an abandoned worker finishing late
cannot report success for a cancelled call.
- clarify (interactive, _NEVER_PARALLEL_TOOLS) keeps the inline path —
it owns its own human wait.
Cooperative layer in the reported offender:
- image_generation_tool: blind handler.get() (generation + Clarity
upscale) replaced with _wait_fal_result(), which polls is_interrupted()
in 0.5s slices and raises ImageGenerationInterrupted immediately.
- _upscale_image propagates the interrupt instead of swallowing it into
the "upscale failed, use original" fallback.
Message alternation is preserved: the cancelled result is a normal tool
result for the call_id. Sabotage-verified: with the old wait loop
restored, the new tests fail (tool blocks full runtime); with the fix
they pass in ~4s.
Shawn Wang