* Release PL (v0.51.451): show named-profile external sessions in the all-profiles list (#4067, #4065)
All-profiles session enumeration now scans across profiles only when the
request explicitly asks (all-profiles view); cross-profile import resolves
to the active profile's state.db for unqualified requests (foreign id -> 404),
mirroring the #3982/#3991 detail/export profile-scoping gate.
Co-authored-by: rodboev
* gate fix (Codex CORE + Opus): close cross-profile leak on the import_cli refresh/existing-session branch
The already-imported branch of _handle_session_import_cli forced allow_all_profiles=True
whenever the globally-stored session carried a profile, with no active-profile visibility
gate — so an unqualified request from profile A could read/refresh a session imported under
profile B's live state.db (both advisors reproduced the foreign-transcript leak). Now gated
exactly like the /api/session detail + export endpoints: unqualified requests require
_session_visible_to_active_profile; explicit all_profiles requires a matching profile.
+ 3 regression tests.
* gate fix (Codex SILENT, re-gate): thread source_filter into the all-profiles CLI scan (models.py:4113)
The all-profiles branch keys its cache on source_filter but called
_load_cli_sessions_uncached without it, so /api/sessions?all_profiles=1 with a
source filter returned every-source sessions under a filtered key. + regression test.
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
* Release PK (v0.51.450): preserve custom provider namespace prefixes in model normalization (#4278)
Token-boundary check in _normalize_provider_id (exact token or prefix
followed strictly by ':' or '/'); explicit aliases for hyphenated
first-party names (google-gemini, google-ai-studio, claude-code); bare-prefix
family detection no longer fires on namespaced model IDs.
Closes#4278. Supersedes #4280 (rodboev) whose '-' boundary re-collapsed
claude-relay->anthropic.
Co-authored-by: b3nw
* gate fix (Codex CORE): add openai-api to _PROVIDER_ALIASES so the :/ token boundary doesn't drop the registered openai-api provider id to '' (would silently repair OpenAI API sessions to the profile default) + 2 regression tests
* gate fix (Opus): gate the streaming send-path bare-prefix loop on '/' not in model so custom namespaces (gemini_cli/, gpt_proxy/, claude-relay/) survive the send path — same #4278 collision Opus found in _apply_profile_provider_context_to_streaming_model; + regression test + CHANGELOG
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
* test-infra: per-process free port + port-keyed state dir (true parallel safety)
ROOT CAUSE of the cascade-under-concurrency: the test port was a fixed
md5(repo_path)%10000, so two concurrent pytest runs from the SAME worktree (e.g.
a dev's suite + a Codex/Opus gate running ./scripts/test.sh to verify) bound the
SAME port — and the session-fixture's _kill_port_owner(TEST_PORT) at setup reaped
the OTHER run's server mid-suite, cascading every HTTP test with ConnectionRefused.
FIX: when HERMES_WEBUI_TEST_PORT isn't pinned, _auto_test_port now grabs a FREE
OS-assigned ephemeral port per process (bind :0, read back, release). The state
dir name is keyed on that port too (webui-test-<repohash>-<port>), so concurrent
runs get fully distinct port + state dir + db — no shared resource, no mutual kill.
VERIFIED: two FULL suites run concurrently from the same worktree → BOTH
9150 passed / 0 failed, 0 ConnectionRefused, 0 dirs leaked into ~/.hermes. This is
the parallel-safe world: Codex + Opus gates + local suite can all run at once.
* test-infra: address Codex CORE findings on the per-process-port fix
1. Don't _kill_port_owner an AUTO (unpinned) port: fuser -k on an ephemeral-range
port can match a client socket or a concurrent run's process, killing the wrong
thing. Track TEST_PORT_PINNED; only reap pinned ports (which may be stale).
Auto ports are freshly confirmed-free, so nothing to reap.
2. Cap auto port at <=64535 so offset-deriving tests (test_server_port_exclusivity
adds up to +905 to TEST_PORT) can't overflow past 65535 (OverflowError).
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
* test-infra: hard-isolate test state from production + robust teardown
Two production-safety fixes to tests/conftest.py + tests/_pytest_port.py:
1. TEST_STATE_DIR now defaults under the OS temp dir (/tmp/hermes-webui-tests/)
instead of HERMES_HOME/webui-test-<hash> — which had been writing test state
INTO the real ~/.hermes/profiles/webui/ (144 leaked webui-test-* dirs found).
Added a hard-stop guard: refuse to run if the resolved test state dir is
inside the production Hermes home, so a misconfigured env can never let tests
wipe/clobber production profiles/sessions/credentials.
2. _rmtree_retry is now resilient on Linux (was a bare shutil.rmtree): retries
transient OSError [Errno 39] Directory-not-empty races (a background daemon
writing into the tree during teardown) + a final ignore_errors sweep, and
warns instead of raising so a benign cleanup race can't spuriously ERROR a
test (the failure mode seen near #4283 context-reconciliation tests).
* test-infra: anchor production-proximity guard on literal ~/.hermes (not $HERMES_HOME)
The first guard compared TEST_STATE_DIR against $HERMES_HOME, which is overridden
to TEST_STATE_DIR itself during a run -> false-tripped collection (6 modules) with
'test dir X is inside production home X'. Now anchor on the literal user-home
~/.hermes and explicitly allow any dir under the OS temp root, so the temp default
never trips while a real ~/.hermes target still hard-stops.
* test-infra: close TMPDIR-under-~/.hermes guard hole (Codex)
Codex flagged: the 'temp is always allowed' exception could be exploited if
TMPDIR itself is set under ~/.hermes — then a 'temp' path is a production path and
_under_temp would wrongly suppress the guard. Now the temp exception only holds
when the temp root is itself OUTSIDE the production home. Verified: normal /tmp
default allowed; TMPDIR=~/.hermes/tmp now trips the guard.
* test-infra: mirror production-proximity guard into _pytest_port.py (Opus defense-in-depth)
* test-infra: remove ~/.hermes production-path fallbacks from 4 test files (Codex)
Codex's kickback found the real gap: test_sprint5, test_sprint45,
test_onboarding_existing_config, test_regressions each hardcoded a
~/.hermes/webui-mvp-test fallback for their state dir (used when running
standalone or if the env var is unset). Repointed all to the shared isolated
TEST_STATE_DIR from _pytest_port (temp-rooted). Also refreshed stale
~/.hermes/webui-mvp-test references in test_gateway_sync comment + ARCHITECTURE.md.
Now ZERO production-path fallbacks remain in tests.
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
test_profile_switch_restarts_watcher_best_effort and
test_profile_switch_response_survives_watcher_restart_failure passed in isolation
but failed in the full suite (HTTP 409 != 200) when an auth-enabling sibling test
left is_auth_enabled() True in-process: /api/profile/switch -> build_profile_cookie
-> sign_profile_cookie_value raises (no session) -> RuntimeError -> 409. Added an
autouse fixture forcing api.auth.is_auth_enabled -> False for this module (these
tests exercise watcher restart, not auth). Verified: full suite 9139 passed, 0
failed. Test-only.
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
* stage-3712: model picker show-all overflow + fix per-row count leak (#3691)
Self-rebased rodboev's #3712 onto master. FIX (Nathan-caught): the per-row
provider chip stripped the '(N of M)' overflow count — that count belongs on the
group heading only, and was reading as nonsense post-expand (row 30 showing
'15 of 30'). Updated test to assert heading-carries-count + rows-do-not.
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
* stage-3712: + dedupe overflow expand (Codex SILENT) — reuse existing injected option
Codex caught: select hidden overflow model from search → reopen → Show all could
duplicate the row (_appendOverflowOptionsToGroup blindly appended every extra
model without checking the select for an already-injected option). Now moves/reuses
an existing same-value option instead of re-creating it.
* stage-3712: + fix expanded-heading double-count (Opus) — strip decoration, count rendered rows
Opus caught: post-expand the heading showed 'Nous (2 of 4) (4)' (appended its own
count onto the already-decorated backend label) and used modelCount (incl hoisted-
configured) vs rendered rows. Now strips the '(N of M)' suffix when expanded +
counts groupRows.length. Test asserts no double count on out[expanded].
* stage-3712: + include extra_models in Settings/Cron/Profile/Aux native selectors (Codex CORE)
Codex caught: the server overflow split moves models>15 into extra_models for ALL
providers, but the 4 native <select> builders in panels.js (settingsModel:6690,
cronFormModel:1112, profileFormModel:5784, _auxProviders:8287) read g.models only
— silently dropping the overflow tail for large providers. Now concat
models+extra_models at all 4 sites. Test guard added.
* Release PG: model picker show-all overflow for large provider catalogs (#3691)
Self-rebased rodboev's #3712 onto v0.51.445; full browser-tested (desktop+mobile,
before/after) + Nathan screenshot-approved. 4 gate-caught issues fixed inline:
per-row count leak (Nathan), overflow-expand dedup (Codex), heading double-count
(Opus), and native Settings/Cron/Profile/Aux selectors dropping the overflow tail
(Codex CORE). Codex SAFE + Opus SHIP + 41 picker tests green.
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
rodboev test-only #4274 (test_issue609, test_sprint3, + 2 others). Linux
assertions stay at full strictness. 63 tests pass. No version stamp (tests-only).
Note: #4273 was NOT included — it touches api/workspace.py runtime (home-expansion
in the blocked-roots boundary), needs a separate full gate.
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
* stage-3629: scope gateway watcher to active profile (profile-keyed registry), rebased onto master
* stage-3629: + stop-race-safe subscribe() (Codex CORE) + regression test
Codex caught a TOCTOU between get_watcher() and subscribe(): a watcher reaped as
subscriber-free during a concurrent profile switch left a late subscriber's queue
without the None sentinel -> SSE loop hangs open (keepalives, no events). Opus
independently found the same race (rated follow-up). subscribe() now checks
_stop_event under _sub_lock after appending and self-enqueues the sentinel.
test_subscribe_after_stop_gets_sentinel_immediately pins it.
* stage-3629: + reconnect gateway SSE to new profile on switch (Codex CORE #2)
Second Codex finding: after a profile switch this tab's _gatewaySSE stayed
subscribed to the PREVIOUS profile's watcher (switchToProfile never restarted it;
the probe reattach is gated on !_gatewaySSE which is false while the old ES is
open). switchToProfile() now calls startGatewaySSE() after S.activeProfile is set
— it closes the old ES + reconnects with the new profile cookie, self-gating on
_showCliSessions.
* Release PF: scope gateway watcher to active profile (#2848)
rodboev's #3629, re-pushed to fix both bounce findings (client SSE reconnect +
singleton→profile-keyed registry). Two Codex CORE races caught + fixed inline:
(1) stop/subscribe sentinel race, (2) tab's own SSE staying on the old profile
after switch. Codex SAFE + Opus SHIP (architecture) + gateway-watcher tests green.
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
Two more rodboev test-only PRs continuing #3664 Windows-harness path
normalization (test_bootstrap_python_selection, test_issue513_wsl_autostart,
test_title_aux_routing). Linux assertions stay at full strictness. Codex SAFE +
full suite 9114 green. No version stamp (tests-only).
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
ai... rodboev's #4133, notification-feature ONLY (dropped the unrelated
'invalidate sidebar cache on WAL' commit — that's a separate good fix to split
out). Button now reflects live Notification.permission instead of silent no-op.
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
* stage-3664: Windows pytest harness hardening (test-only batch #4254/4255/4256/4257/4259)
* test(#3664): Windows pytest-harness compatibility batch
Combines 5 rodboev test-only PRs (#4254/#4255/#4256/#4257/#4259), all under tests/.
Linux assertion paths preserved at full strictness; relaxations + new markers are
Windows/no-fork gated. Codex SAFE + Opus SHIP + changed tests green (cascade-only
full-suite noise is the box artifact). No version stamp (tests-only, Nathan rule).
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
* stage-3955: blur composer on Escape for j/k nav + IME guard (Opus nit)
* stage-3955: update test to match IME-guarded blur branch
* Release OX: blur composer on Escape for j/k message nav (#3952)
Self-rebased TomBanksAU's #3955 onto v0.51.436; full Codex+Opus+suite gate.
Escape now blurs the focused composer so j/k session-nav works, after higher-
priority Escape actions run; dropdown-Escape stops propagation; IME-composition
guarded (Opus nit). Also restored the v0.51.434/435 CHANGELOG headings that an
earlier batch left folded under v0.51.436 (hygiene fix).
Co-authored-by: TomBanksAU <TomBanksAU@users.noreply.github.com>
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: TomBanksAU <TomBanksAU@users.noreply.github.com>
* stage-3921: cron-session unread disambiguation (#3460, rodboev)
* Release OW: cron sessions mark unread on new runs (#3460)
Self-rebased rodboev's #3921 onto v0.51.435; full Codex+Opus+suite gate.
Fixes cron-session unread cross-attribution: resolver now considers all known
job ids and attributes each session to the longest matching cron_<jobid>_ prefix
(provably unambiguous). Read-only state.db scan, fires only on real completions.
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
* fix(#4164): bound project-context file walk in non-git workspaces
_project_context_candidates mirrored the agent's _find_hermes_md exactly,
including the bug that 'stop_at = git_root' is None in a non-git tree —
so the walk 'for directory in [cwd, *cwd.parents]: ... if stop_at and
directory == stop_at: break' never broke and ran to filesystem root.
A workspace at /tmp/x/project/subdir could surface /tmp/x/HERMES.md
(a file outside the workspace) in the Memory > Project Context tab.
Fix: when no git root is found, treat the workspace itself as the stop
boundary. The cwd is still scanned (preserving in-workspace AGENTS.md /
HERMES.md behavior), but parents are skipped — we never walk into the
user's home directory or /tmp.
Git workspaces are unchanged (still walk up to the git root). The
existing test_3866 walk-to-git-root test still passes.
The issue notes the agent-side walk in agent/prompt_builder._find_hermes_md
should be bounded the same way for full parity; that lives in a separate
repo (hermes-agent), so this PR fixes only the WebUI side. Until the
agent ships the same bound, the WebUI tab will under-report context
files that live above a non-git workspace — strictly less surprising
than the previous over-report.
tests/test_issue4164_bound_non_git_project_context_walk.py (4 tests) pins:
- non-git workspace does not leak a parent HERMES.md
- non-git workspace still reads in-workspace context files
- git workspace walk-to-git-root is unchanged
- workspace-level HERMES.md wins, parent copy doesn't appear even shadowed
All 10 existing test_3866_project_context_memory_section tests still pass.
Closes#4164.
* chore(release): stamp v0.51.428 (Release OO) for #4164
---------
Co-authored-by: Sanjay Santhanam <51058514+Sanjays2402@users.noreply.github.com>
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
* fix(#3988): default show_cli_sessions to true for new installs, grandfather existing OFF
Combines the best of #3997 (rodboev) and #4222 (Sanjays2402):
- config.py default flipped True so NEW installs surface CLI/TUI/Telegram/
Discord/Hermes One sessions in the sidebar without hunting for the toggle.
- load_settings() backfill grandfathers ESTABLISHED installs OFF: when the saved
settings predate the flip (onboarding_completed True + key absent), pin False so
an existing user's sidebar doesn't silently change. Fresh installs get True.
- Default-true hydration at every read site (config.py / boot.js / panels.js
checkbox) via the !==false idiom — no config-ON-but-UI-OFF mismatch (#4006 class).
- 10 tests: config default, bool-key round-trip, grandfather matrix
(fresh/mid-onboarding/established-absent/explicit-true/explicit-false/
save-preserves-grandfather), and both frontend hydration sites.
Co-authored-by: Rod Boev <rod.boev@gmail.com>
Co-authored-by: Sanjays2402 <Sanjays2402@users.noreply.github.com>
* fix(#4217): guard /api/file/delete and /api/file/rename against symlink targets
* chore(release): stamp v0.51.427 (Release ON) for #3988 + #4217
* chore: drop unused imports (ruff F401) in absorbed test files
* fix(#4217,#3988): address review — dangling-symlink ordering + boot.js catch fallback
Greptile review on #4230 + the related-PR reviews surfaced two fixes on the
shipped code:
1. (#4217) _handle_file_delete/_handle_file_rename checked target.exists()
BEFORE is_symlink(), so a DANGLING workspace symlink resolved to a missing
target -> misclassified as 404 'File not found' and left permanently
undeletable/unrenameable. Moved the no-follow is_symlink() guard before the
follow-based exists() check in both handlers. +2 dangling-symlink tests.
2. (#3988) boot.js settings-load-FAILURE catch block hardcoded
_showCliSessions=false, inconsistent with the new True default (the #4006
catch-block-fallback class). Now defaults true. +1 test pinning the fallback.
Both confirmed by the new tests (20/20 in the two files).
* fix(#3988,#4217): address Opus review — broaden grandfather signal + move-handler parity
Opus #4230 ship-review surfaced 2 LOW findings, both now fixed in-line:
1. (#3988) The grandfather keyed only on onboarding_completed, so a CLI-configured
user who tweaked a WebUI setting but never ran the wizard would get flipped ON
(and onboarding.py later auto-persisting onboarding_completed=True would make it
permanent). Now: established = show_cli_sessions absent AND (onboarding_completed
OR any persisted setting other than a falsy onboarding flag). A genuinely new /
still-mid-onboarding file still falls through to True. +1 test.
2. (#4217) _handle_file_move still checked exists() before is_symlink() — the same
dangling-symlink-404 bug just fixed in delete/rename. Reordered for parity so a
dangling symlinked source is rejected 400, not 404.
22 tests across the two files pass; ruff clean.
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: Rod Boev <rod.boev@gmail.com>
Co-authored-by: Sanjays2402 <Sanjays2402@users.noreply.github.com>
* fix: keep custom provider when model id has known-provider prefix (#4210)
Siblings of #3872 (v0.51.349, custom provider with a configured base_url
preserves intrinsic routing prefixes). The base_url unset variant of
the same bug was missed: with provider=custom or custom:<slug> and no
base_url set, a model id like 'x-ai/grok-2' or 'google/gemma-2' was
hijacked to provider='openrouter', base_url=None. The backend then
failed to initialize with 'RuntimeError: No LLM provider configured'
because the user had no OpenRouter credentials configured.
resolve_model_provider() now applies the same _is_custom exemption the
config_base_url branch already carries to the cross-provider redirect
at the end of the slash branch. The custom provider keeps the request
and the full model id; real first-party providers pointed at an
OpenAI-compatible proxy (e.g. provider=openai + base_url=litellm) are
unaffected because their strip path is in the base_url branch and the
'prefix == config_provider' branch above the change.
Verified:
- test_model_resolver.py 32 passed (30 existing + 2 new RED tests)
- test_custom_provider_bare_model_reasoning.py 34 passed
- Pre-existing F401 (api/config.py:23,24,28) are not introduced by
this change; same 3 errors exist on upstream/master.
Refs #4210
* chore(release): reposition #4210 CHANGELOG entry under v0.51.426 (OM)
---------
Co-authored-by: b3nw <b3nw@users.noreply.github.com>
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
* fix: bound paginated session tails by visible messages
* fix(#4070): always merge state.db tail rows (drop the stale-tail timestamp shortcut)
Deep-review finding: the sidecar-only timestamp shortcut returned the sidecar
verbatim whenever state.db had no strictly-newer rows, but a state.db row
at-or-before the sidecar's newest timestamp (recovery / edited-in-place /
missing-timestamp) is still absent from the sidecar -> the paginated tail
diverged from the full merge path (empirically 3/6 adversarial cases). Always
run the cheap in-memory append-only merge; the real latency win (skipping the
lineage-parent DISK load) is preserved.
Co-authored-by: t3chn0pr13st <t3chn0pr13st@users.noreply.github.com>
* fix(#4070): keep matching trailing tool-result row in paginated window
Codex ship-gate finding: _message_window_for_display pulled end_idx back to the
last renderable (user/assistant) row, dropping a trailing role:tool result row
even when it's the result of the newest assistant tool-call. The renderer
rebuilds tool cards from these rows (CLI-origin / empty S.toolCalls path,
static/ui.js resultsByTid), so the card lost its result snippet. Now extend
end_idx forward through trailing tool rows whose tool_call_id matches a tool-call
present in the window; orphan trailing tool rows (no match) are still skipped to
preserve the visible-row budget. +2 regression tests.
Co-authored-by: t3chn0pr13st <t3chn0pr13st@users.noreply.github.com>
* chore(#4070): remove now-dead _max_message_timestamp + its import
Opus gate nit: _max_message_timestamp was only called by the removed
sidecar-timestamp shortcut, and the _message_timestamp_as_float import existed
solely to feed it. Both removed; _message_timestamp_as_float still lives in
api/models.py for its real callers.
* chore: add #4069 CHANGELOG entry (visible-message tail pagination)
* chore(release): stamp v0.51.424 (Release OK) for #4069
---------
Co-authored-by: t3chn0pr13st <technopriest@live.ru>
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: t3chn0pr13st <t3chn0pr13st@users.noreply.github.com>
* perf(chat): virtualize long message transcripts (#500)
* fix(chat): preserve virtual height cache on append (#500)
* fix(chat): bound virtual tail follow-up (#500)
* fix(chat): surface older history and bound virtual rerenders (#500)
* fix(chat): stabilize virtual history fallback (#500)
* fix(chat): preserve virtual compression anchors (#500)
* Keep virtualization regressions aligned with rebased sources (#500)
* fix(chat): preserve virtual heights across prepends (#500)
* test(update): isolate restart timing checks (#814)
* fix(chat): restore render contracts around virtual cache
* fix(chat): preserve legacy assistant anchor predicates
* fix(#3714): restore worklog-disclosure capture after HTML-cache early-return
Deep-review finding: the capture of _captureWorklogDetailDisclosureState was
moved before the HTML-cache fast-path return, failing the invariant test
test_render_rebuild_preserves_worklog_detail_disclosure_click_state (it must
not traverse the previous-session DOM when the cache path returns early).
Moved it back to just before inner.innerHTML='' (after the cache return).
Co-authored-by: Rod Boev <rod.boev@gmail.com>
* fix(#3714): restore Transparent Stream rehydration in HTML-cache fast path
Full-suite gate caught test_issue3820_chat_activity_display_mode failure: the
virtualization rewrite of the renderMessages HTML-cache fast-path block dropped
the _rehydrateTransparentStreamDom(inner) call (master has it; PR had 2 of 3
occurrences). On a cache-restore session switch, Transparent Stream controls
lost their property-assigned handlers. Restored the call in the same position
master has it (right after _sessionHtmlCacheSid=sid).
Co-authored-by: Rod Boev <rod.boev@gmail.com>
* fix(#3714): preserve live turn on scroll-rebuild even with zero streamed text
Opus ship-review residual (now fixed, not deferred): the live-turn re-attach
after a scroll-triggered virtualized rebuild was gated on _preservedLen>0, so a
turn showing only Activity/tool/worklog blocks with no streamed text yet (mid
tool-call) would not be preserved — scrolling up into history on a >80-row
transcript mid-stream could blink those live-only blocks for a frame. Broaden
the gate to also restore when the preserved turn has MORE structural blocks than
the rebuilt (lagging-S.messages) turn. Hoisted the _structuralCount helper above
the gate; whole-turn-vs-segment restore logic unchanged.
Co-authored-by: Rod Boev <rod.boev@gmail.com>
* chore(release): stamp v0.51.423 (Release OJ) for #500 transcript virtualization
---------
Co-authored-by: Rod Boev <rod.boev@gmail.com>
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
* feat(workspace): add Todos tab to workspace panel behind a setting (#2942)
* fix(workspace): use DOM attribute to check active tab across script scope
* Prefer the live todos snapshot in the workspace tab
* Fill the missing workspace Todos locale keys
* feat(workspace): add optional Todos tab to workspace panel (#3564)
Rebased rodboev's #3564 onto current master + CHANGELOG entry.
Co-authored-by: Rod Boev <rod.boev@gmail.com>
* fix(#3564): apply workspace Todos tab visibility on cold boot
Codex gate: boot.js set window._workspaceTodosTab from saved settings but
never called _applyWorkspaceTodosTabVisibility(), so a previously-enabled
Todos tab stayed hidden on a normal page reload (only re-appeared after
opening Settings). Call the visibility helper after assigning the flag in
both the settings-apply and the fallback/default paths.
Co-authored-by: Rod Boev <rod.boev@gmail.com>
* chore(release): stamp v0.51.422 (Release OI) for #3564
---------
Co-authored-by: Rod Boev <rod.boev@gmail.com>
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
* fix(#3750): preserve LM Studio auth on reasoning probe
Rebased rodboev's #3837 onto current master (master reworked
resolve_model_reasoning_efforts via #3431 with a combined hermes_cli
import + _nested_route_reasoning_denied guard). Threaded the api_key
resolution + the _lmstudio_model_reasoning_options wrapper + the
HTTP-fallback probe into master's current structure; intent preserved
byte-identically (+164/-24, same as the original PR diff).
Co-authored-by: Rod Boev <rod.boev@gmail.com>
* fix(#3837): gate LM Studio probe credential to configured endpoint + no-redirect opener
Codex security review (CORE) on the rebased #3837: GET /api/reasoning accepts
a caller-supplied base_url, and the new probe attached the configured LM Studio
API key as a Bearer token; urllib also re-sends Authorization across redirects.
Together that could exfiltrate the stored credential to an attacker-controlled
host. (master only escalates here because master probed WITHOUT a key.)
Fixes:
1. Only forward the LM Studio credential when the probe target normalizes to
the configured LM Studio base URL (reuses _normalize_base_url_for_match);
a caller-supplied non-matching base_url is probed keyless.
2. _lmstudio_reasoning_probe_options_fallback uses a no-redirect opener
(_NoRedirectHandler) so a 3xx can't forward the Authorization header onward.
3. Two regression tests: caller-supplied base_url drops the key; redirect is
not followed (credential never reaches the redirect target).
Co-authored-by: Rod Boev <rod.boev@gmail.com>
* fix(#3837): route ALL credentialed LM Studio probes through the no-redirect probe
Codex re-gate caught that the first fix only hardened the built-in fallback;
in production hermes_cli IS importable, so the preferred path called
hermes_cli.models.lmstudio_model_reasoning_options(..., api_key=...) which uses
plain urlopen and still follows redirects with the credential.
Fix: _lmstudio_model_reasoning_options now short-circuits to the built-in
no-redirect probe whenever api_key is set; the hermes_cli path is used ONLY for
keyless probes (no credential to leak). Reworked tests accordingly:
- test_credentialed_probe_never_calls_hermes_cli: credentialed probe bypasses CLI
- test_keyless_probe_logs_signature_mismatch_before_fallback: keyless CLI degrade
- test_credentialed_probe_does_not_follow_redirects_end_to_end: E2E through
resolve_model_reasoning_efforts with real hermes_cli; redirect target never hit
Verified both security tests fail without the guard and pass with it.
Co-authored-by: Rod Boev <rod.boev@gmail.com>
* chore(release): stamp v0.51.420 (Release OG) in CHANGELOG
---------
Co-authored-by: nesquena-hermes <agent@nesquena-hermes>
Co-authored-by: Rod Boev <rod.boev@gmail.com>
Combined light-UX release NW. Both rebased onto current master (were 96 and
247 behind; net 3-dot diffs applied, byte-identical to PR heads). Both reuse
existing renderers (chat image lightbox / chat CSV table renderer).
Review fix (#4025, Codex): the CSV table preview previously stripped the
workspace Edit affordance .csv had when it fell through to the code preview.
Restored: csv mode is editable, renderCsvPreviewContent caches raw content for
the edit textarea, and saving an edited CSV re-renders the table (not markdown).
Verified the full edit round-trip live + regression test.
Co-authored-by: Rod Boev <rod.boev@gmail.com>
Consolidates #4162 + #3944. Aux path: skip the reasoning-disable inject for
routes that 400 on it via _route_rejects_reasoning_extra — hostname-matched
OpenAI / Azure OpenAI / Azure AI Foundry (+ provider aliases openai*/azure*/
azure-foundry/azure-ai*) and OpenRouter Anthropic mandatory-reasoning; keep it
for local/OpenRouter-non-anthropic/other reasoning routes (#4161). Agent path:
rely on the existing agent.reasoning_config={enabled:False} + _build_api_kwargs()
(route-correct per provider profile) rather than re-injecting a generic disable
on top of it; only MiniMax reasoning_split is added (unchanged from master).
Co-authored-by: Rod Boev <rod.boev@gmail.com>
Re-pushed by rodboev fixing both bounce findings (parent streaming-state
isolation + nested-fork batch-select) plus a select-handler selector fix.
On current master (0 behind); sessions.js byte-identical to PR head.
Co-authored-by: Rod Boev <rod.boev@gmail.com>
- Return None for blank session.workspace instead of Path('').resolve()
(which resolved to the server CWD and could surface the install's own
AGENTS.md/HERMES.md as project context).
- Match the agent's lowercase filename variants (agents.md / claude.md).
- Read .cursor/rules/*.mdc, not just the legacy .cursorrules file.
- Strip YAML frontmatter and cap each source at 20k, mirroring what the
agent actually injects.
- Regression tests for the guard, lowercase match, and frontmatter strip.
In Transparent Stream mode a live multi-round turn (narrate -> tools ->
narrate -> tools) rendered all assistant prose bunched at the top with
every tool/thinking row clustered below; it self-healed only once the
turn settled.
_syncLiveWorklogReasonsForAnchor() runs on every live segment render
(_flushPendingSegmentRender + the RAF _doRender in messages.js) and
unconditionally built the top-anchored live-worklog rail, mirrored each
round's prose into a wl-reason row there, and hid the real inline
assistant-segment via assistant-segment-worklog-source (display:none).
That prose-folding is the Compact Worklog presentation (#3401) and must
not run in Transparent Stream mode.
Gate the helper on isCompactWorklogMode() so transparent mode keeps prose
as visible, chronologically-placed inline segments interleaved with tool
rows (matching the already-correct settled render). Compact Worklog
folding is unchanged. Verified in an isolated mock-LLM env via an
on-screen visual-position probe (before: textAfterFirstTool=0,
allTextAboveAllTools=true; after: textAfterFirstTool=3,
allTextAboveAllTools=false) and confirmed compact_worklog still folds.
Adds a behavioral regression test asserting the gate is the first
statement in the helper (precedes the rail build), red/green validated.
The #4151 focus hook reopened both global sidebar streams on `focus` with a
bare `startGatewaySSE()`. But `startGatewaySSE()` is NOT idempotent — it begins
with an unconditional `stopGatewaySSE()` (sessions.js:3778), which closes the
live gateway EventSource, cancels its poll fallback (`stopGatewayPollFallback`),
and resets `_gatewayProbeInFlight` / `_gatewaySSEWarningShown`.
So on a transient blur shorter than the 1s debounce — where the blur-close
timer is cleared on the following focus and the stream was never torn down —
the focus handler still dropped and reconnected the gateway on every
window-switch, cancelling its fallback and resetting probe/warning state. That
is the exact thrash the debounce exists to prevent, in the multi-window PWA
scenario this PR targets. (`ensureSessionEventsSSE()` is already idempotent via
`if(_sessionEventsSSE) return`, so the session-events stream was unaffected;
only the gateway churned. greptile P1.)
Guard the reopen on the gateway actually being closed: `if(!_gatewaySSE)
startGatewaySSE()`, mirroring the session-events idempotency. A sustained blur
(timer fires → both streams closed) still reopens the gateway on focus; a
transient blur leaves the still-open gateway untouched. Verified with a
blur/focus/timer harness (transient → no teardown/reconnect; sustained → close
then reopen; CLI-disabled → stays closed) and a new regression guard
`test_focus_reopen_does_not_thrash_the_gateway_stream`.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Two side-by-side PWA standalone windows saturate the browser's per-origin
HTTP/1.1 connection limit, causing 'Request timed out' toasts to cycle every
~30s. The idle-SSE close added in #3992/#3996 keys off the Page Visibility API
(visibilitychange / document.hidden), but a PWA standalone window does not
reliably fire visibilitychange when it loses focus to another window of the
same app -- document.hidden only flips on minimize. So both windows stay
'visible', each holds its two global sidebar SSE streams (session-events +
gateway) open, 2x3 = 6 hits the connection cap, and subsequent fetch() calls
(the 30s background polls) queue behind the saturated pool and time out.
The two global sidebar streams now also close on a sustained window blur (gated
on document.hasFocus(), the signal visibilitychange misses) and reopen --
catching up the session list -- on focus, via a shared _sidebarSseBackgrounded()
predicate and a debounced _installSidebarSseFocusHook().
The per-session live stream (messages.js startSessionStream) is intentionally
left visibility-only: it carries live bg_task_complete toasts and
server_turn_started live-view that an unfocused-but-visible window must still
receive (the OS-notification path is gated on document.hidden, so the in-app
toast is the only completion signal a visible-unfocused window gets). Closing
it on blur would regress the multi-window live-view UX.
Reported by @ddiall with a thorough root-cause analysis.
Co-authored-by: ddiall <ddiall@users.noreply.github.com>
get_reasoning_status() returned the raw config reasoning_effort, so a stale saved
'max' surfaced on boot/status/chip even though streaming coerces it to xhigh. Now
routes the status value through coerce_reasoning_effort_for_model() with the resolved
model/provider/base_url, so all read paths agree. Added a stale-max->xhigh status
regression test.
#4124: treat preserved snapshot bottom>250px as active reader (restore before follow-to-bottom),
re-mark restored mid-answer as unpinned, scope _resetScrollDirectionTracker to real session
switches only. Re-anchored test_preserve_scroll_restores_unpinned_viewport + test_render_messages_
preserve_scroll to the new readerAwayFromBottom-guarded early-return (was pinning the old exact line).
CHANGELOG stamped v0.51.390 (NC).
rfind for the Phase-2 INFLIGHT branch (idle-reset added an earlier if(INFLIGHT[sid]){);
and test_load_session_attaches_sse_before_auxiliary_work now asserts each auxiliary call
individually (the old contiguous syncTopbar();renderMessages();appendThinking();loadDir
literal was legitimately broken up by the new restoreLiveTurnHtmlForSession insertion +
renderMessages(preserveScroll) arg — all calls still run after attachLiveStream). Behavior unchanged.
test_issue3306 _load_session_clear_block (4000->6500) + test_issue_new_chat_draft_restore
body window (4000->6500): #3899's idle-reset + live-turn-snapshot blocks pushed the
carry-forward snapshot + destructive S.messages clear past the old fixed windows. Behavior
+ ordering unchanged; windows widened to contain them.
#3899's new idle-reset added an EARLIER if(INFLIGHT[sid]){ block in loadSession, so
4 test_regressions tests' src.find('if(INFLIGHT[sid]){') grabbed the wrong (idle-cleanup)
occurrence instead of the Phase-2 restore branch → switched to rfind (the substantive
branch). And the idle-cleanup comment changed from 'discard any browser-side inflight'
to 'reset browser-side streaming flags' → re-anchored test_frontend_drops_inflight_cache
on the new comment + nested if(!activeStreamId){...if(INFLIGHT[sid]){ form (behavior
preserved + enhanced). No production code changed.
Production fixes (Tamaz-sujashvili, reviewed sound by maintainer):
- loadSession clears S.busy/S.activeStreamId as soon as metadata confirms no active_stream_id, before the async message-load gap (idle session no longer shows streaming chrome).
- Snapshots the live turn before wiping msgInner + seeds INFLIGHT, restores on the active-stream return path (timer/trace survive switch-back).
Re-anchored the 2 brittle regression tests per maintainer review: test_..snapshots.. now anchors on the unique 'Loading conversation...' marker (was matching the no-space 'Session not available' error path); test_..restores.. now asserts the LIVE Phase 2a restore (after loadInflightState) instead of the unreachable Phase-2b/1184 branch. CHANGELOG stamped v0.51.384 (MW).
The update-check payload now carries a `dirty: bool` reflecting the
working-tree state vs HEAD, computed via the same `git diff-index
--quiet HEAD --` primitive already used by `_dirty_suffix()`. A dirty
install at-or-past the latest release tag used to silently report
"Up to date" with no remediation affordance; the Settings panel
can now read this flag to offer `apply_force_update` (which already
does `git checkout .` + `git reset --hard` to reset to a clean
latest).
The probe is conservatively clean on real errors (timeout, missing
git, fatal) so a transient probe failure never produces a false-
positive "local changes" alert.
Closes#4085
scrollToBottom now calls _settleMessageScrollToBottom(false, true); renderMessages
captures scrollSnapshot when (preserveScroll || (!_autoScrollFollow && _messageUserUnpinned)).
Updated the 4 change-detector assertions in test_tars_scroll_reset_regressions.py +
test_issue1690_scroll_completion.py to the new strings; behavioral intent unchanged.
Per maintainer decision, auto_scroll_follow defaults TRUE (sticky-bottom: follow
new output while streaming, scroll-up unpins and is respected) — matching Codex
CLI / Claude Code. The PR shipped it default-False.
Default-mismatch bug fixed (would have shown the True default as OFF for existing
users with no saved value):
- boot.js settings path: !!s.auto_scroll_follow -> s.auto_scroll_follow!==false
- boot.js no-settings fallback: false -> true
- panels.js checkbox render: !!settings.auto_scroll_follow -> !==false
Resolved the config.py + panels.js merge conflicts (kept both the Transparent
Stream setting and auto_scroll_follow). Added test_issue4006_auto_scroll_follow_default.py
pinning the True default across config/boot/panels + the follow-gate references.
The 2 failing assertions pinned the removed implementation (const passes=[0,16,80,180]
setTimeout fan-out + _settleMessageScrollToBottom(true) in scrollToBottom). Updated to
assert the new behavior contract: ResizeObserver on #msgInner + 2s static fallback +
force=false so the observer runs. Behavioral intent (settle survives late markdown
growth; sync write before settle scheduling) preserved.
Opus final gate caught that the staged re-implementation placed the
window._showThinking===false gate in _worklogReasonNodeFromText, which has
ZERO callers — so live/settled reasoning rows (built by _syncWorklogReasonFromAnchor
and _appendWorklogReason) were never actually hidden, and removeThinking swept
data-worklog-reason-source=reasoning which the real rows (data-worklog-anchor-reason=1)
don't carry. The string-match tests passed against the dead function = green-suite blind spot.
Fix:
- Gate _syncWorklogReasonFromAnchor (live + settled): when _showThinking===false, remove any
existing reasoning row and bail before building.
- Gate _appendWorklogReason (settled rebuild): return null when _showThinking===false.
- removeThinking + the settled-rebuild cleanup now sweep BOTH data-worklog-anchor-reason=1
(the real attribute) and the legacy data-worklog-reason-source=reasoning.
- Re-anchor the tests to assert against the real render functions + the correct attribute.
_active_skill_search_dirs filters to existing dirs, so on a host with no local
skills dir but configured external dirs the local root is dropped from the list
and the position-based skills_dirs[0]==local assumption misclassified the first
external root as local (its flat skills silently lost their category label).
Pass the local dir explicitly (backward-compatible optional param defaulting to
the old skills_dirs[0] behavior) + regression test for the absent-local-dir case.
The richer static-catalog builder's group-building loop used a strict item["id"]
extraction that would KeyError (caught → degrade to the minimal one-model catalog)
for legal config shapes where providers.<id>.models is a list of dicts keyed by
"model"/"name" rather than "id". Mirror the tolerant id-or-model-or-name resolution
the detection loop already uses, skip entries with nothing usable, and add a
regression test. No behavior change for the common list-of-strings / list-of-id-dicts
shapes.
docs(changelog): stamp #4052 model-picker budget fallback as v0.51.370 (Release MI)
Resolve custom provider API keys from the matched config snapshot and pass them through session hydration plus streaming fallback context-length probes. This prevents authenticated /v1/models endpoints from falling back to the default 256K window and clobbering larger persisted session metadata.
Adds 4 fr-CA + 3 fr-FR Edge neural voices so francophone users can use
the Edge TTS engine instead of receiving HTTP 400 "invalid voice" on
every utterance. Pure superset of the existing allowlist; no validation
or rate-limit behavior changes.
Tests mirror the test_issue2931 in-process / mocked-edge_tts pattern:
each new voice is parametrized through _handle_tts and asserted to
reach synthesis (HTTP 200); fr-BE-CharlineNeural (real Edge voice
but intentionally unlisted) is asserted to still 400.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The profile cookie has been configurable since #1756 via
WEBUI_PROFILE_COOKIE_NAME, the lone WebUI env var missing the HERMES_WEBUI_
prefix shared by every other setting (e.g. HERMES_WEBUI_COOKIE_NAME from #3981).
- Read HERMES_WEBUI_PROFILE_COOKIE_NAME first (canonical name)
- Keep WEBUI_PROFILE_COOKIE_NAME as a deprecated alias so existing deployments
are unaffected; behavior is unchanged, only the name is aligned
- Warn once per process for the legacy name (this resolver runs on every
request, so the deprecation log must not fire per-request)
- Add resolution tests covering canonical, legacy, precedence, blank, and
warn-once paths
#4022 now calls resolve_runtime_provider(requested=..., target_model=...) so
per-model base_url normalization applies; the sprint42 test pinned the old
single-arg signature. Updated the assertion to expect target_model (pre-existing
test, not a behavior regression — confirmed it passes on master only because the
call signature changed here).
Codex caught a regression: the contributor's _runtime_preferred_base_url preferred
the runtime base_url for ALL non-custom providers, which would clobber an explicit
providers.<id>.base_url override pointing at a different host/port (LM Studio at a
LAN IP, an OpenRouter mirror). Now only prefer the runtime URL when it's the SAME
scheme+host+port as the configured one (the #3895 /v1-dedup case is path-only on
the same host); a genuinely different configured endpoint wins. Added
_same_base_url_endpoint() + 2 regression tests (different-endpoint preserved,
same-endpoint normalization).
Opus independent security review concurred SAFE and surfaced 2 LOW defense-in-depth
items, both applied: (1) verify_profile_cookie_value now validates the profile name
against _PROFILE_ID_RE itself (not only in get_profile_cookie) so a future second
caller can't return an unvalidated name; (2) build_profile_cookie raises when auth is
enabled and no handler is passed, so a future call site can't silently emit an
unsigned (session-unbound) profile cookie. +3 regression tests.
The stale-load guard (#3993 Codex race fix) adds an early at the top
of the loadSession catch block; the test helper anchored on the first
after the catch, which now truncated at the guard. Anchor the body
extraction past the guard (from ) so it captures the 404/non-404
self-heal body as before. Behavior unchanged.
currentSid is snapshotted before the awaited /api/session fetch; if the user
clicks a healthy session while a boot-time restore is in flight and that boot
load then fails non-404, _clearStuckSessionOnBoot(sid, null) would wipe the
healthy session's localStorage/URL. Guard the catch block on
_loadingSessionId !== sid (a newer load superseded this one) — re-arm the active
stream and bail before any self-heal/DOM mutation. Protects both the non-404 and
404 inline self-heal paths. + regression test.
Refined the Codex-CORE fix: the happy-path (already-stored session) guard rejects
only an explicit read_only flag — a stored messaging session already owns its
sidecar, so the messaging-fork risk is specific to the materialize FALLBACK
(which creates a sidecar) where the _is_messaging_session_record check stays.
Also completed test_issue1436's _stub_session MagicMock (read_only=False,
_loaded_metadata_only=False) — bare MagicMock auto-attrs were truthy, tripping
the new read-only guard + _ensure_full_session_before_mutation reload.
Codex caught two data-integrity gaps in the contributor's guard: (1) the
get_session() happy path returned a stored session without checking read_only/
messaging, so an already-imported read-only session could be mutated via
rename/update/move; (2) the fallback only checked cli_meta.read_only, but agent
rows normalize messaging sources WITHOUT setting read_only — materializing a
writable sidecar for a state.db-owned messaging session forks its title/state.
Now reject getattr(s,'read_only') OR _is_messaging_session_record on the happy
path, and cli_meta.read_only OR _is_messaging_session_record(cli_meta) in the
fallback. Replaced the messaging-stub test with 3 regression tests (stored
read-only, stored messaging, messaging cli_meta without read_only flag).
Two scroll-state bugs cause the viewport to snap to the bottom after
streaming completes, even when the user has scrolled up to read:
1. _scrollAfterMessageRender (ui.js): when preserveScroll=true and
_scrollPinned=true (re-set by the 250px near-bottom hysteresis),
_followMessagesAfterDomReplace() calls scrollToBottom(), overriding
the user's position. Added !_messageUserUnpinned guard so explicitly
scrolled-up users get their snapshot restored instead.
2. _finishDone (messages.js): the explicit scrollToBottom() at L3302
fires when shouldFollowOnDone is true, but that flag only checks a
120px near-bottom threshold. Added _isMessagePaneNearBottom(250)
gate so users reading 250px+ above bottom keep their position.
Also adds a new 'auto_scroll_follow' setting (default: off) that gates
scrollIfPinned() and _shouldFollowMessagesOnDomReplace(). When disabled,
the viewport never auto-scrolls during streaming — the user controls
scroll position manually and uses the ↓ button to jump to bottom.
7-file setting-add pattern applied:
- api/config.py: _SETTINGS_DEFAULTS + _SETTINGS_BOOL_KEYS
- static/boot.js: both init paths
- static/ui.js: scrollIfPinned() + _shouldFollowMessagesOnDomReplace()
- static/index.html: checkbox in appearance settings
- static/i18n.js: en + zh-CN translations
- static/panels.js: payload, autosave, populate, full save
- messages.js: no additional changes (gated via _shouldFollowMessagesOnDomReplace)
Tests updated: test_tars_scroll_reset_regressions, test_issue1690,
test_issue3545. All 116 scroll-related tests pass.
If startSessionStream(sid) runs while the tab is ALREADY hidden (session loaded/
restored in a background tab), it set _sessionStreamSessionId then returned at the
hidden-skip WITHOUT recording _sessionStreamHiddenSid — so the visibility handler
(which reopens only from the holder) never reattached, silently dropping
bg_task_complete/server_turn_started until the next explicit startSessionStream.
Set _sessionStreamHiddenSid = sid on the hidden-skip return. Adds a regression test.
Add HERMES_WEBUI_COOKIE_NAME so multiple WebUI instances sharing a
hostname (different ports) can use distinct auth cookie names. Browsers
scope cookies by host, not host+port (RFC 6265), so same-host instances
otherwise trample each other's `hermes_session` cookie and log users out.
- Resolve the cookie name via _resolve_cookie_name(): env > default,
mirroring the existing _resolve_session_ttl() pattern.
- Keep `hermes_session` as the default for backwards compatibility.
- Validate against the RFC 6265 token grammar; fall back to the default
with a logged warning on empty or malformed values.
- Cover default, override, empty, invalid, and Set-Cookie paths in
tests/test_auth_sessions.py::TestCookieNameResolution.
#3964 [security] gate first-password bootstrap (_set_password on POST /api/settings
while auth disabled) to local clients — blocks remote unauth first-run ownership.
Uses request-start auth snapshot (auth_enabled_before), so no mid-request TOCTOU.
Self-rebased onto v0.51.357 (8-behind, 3-dot fidelity verified byte-identical).
#3970 (oauth single-flight) DROPPED from this stage: Codex+Opus both caught a
check-then-insert race — _pending_oauth_flow_for releases the lock before the
device-code request + flow insertion, so concurrent unauth starts still spawn
multiple workers (Codex empirically reproduced w/ 2 threads). Returned to author
w/ the atomic per-(provider,home) start-lock fix. Re-gating the (N-1) stage.
Co-authored-by: Hinotoi-agent <Hinotoi-agent@users.noreply.github.com>
