Tool-schema assembly at CLI/Desktop startup runs the browser-family
check_fns (browser, browser_cdp, browser_dialog, browser_vision). Each
of those gates called _get_cdp_override(), which resolves the configured
endpoint over HTTP (GET /json/version, timeout=10) — so a *stale*
browser.cdp_url pointing at a dead debug browser cost ~7 serial blocking
socket connects before the banner rendered. Measured on a real Windows
install with a dead http://[::1]:9222 config: 15.1s of an 18s launch,
with no warning or error — just mystery slowness. The value is easy to
leave behind: /browser connect writes a session-scoped env override, but
'hermes config set browser.cdp_url' persists forever while the debug
Chrome it pointed at dies on the next browser restart.
Split the helper:
- _get_cdp_override_raw() — returns the configured value (env var or
config.yaml) with zero network I/O. Used by every is-it-configured
gate: check_browser_requirements, _browser_cdp_check, _is_local_mode,
_is_local_backend, _navigation_session_key, _should_inject_engine
(via _is_local_mode), and the hermes doctor chromium-skip check.
- _get_cdp_override() — unchanged contract (raw + /json/version
resolution), now only called on paths that are about to connect:
session creation and the dialog-supervisor attach.
This follows the existing rule in check_browser_requirements ('do not
execute agent-browser --version here') and the browser.manage status
path, which already banned _get_cdp_override for exactly this reason
(test_browser_manage_status_does_not_call_get_cdp_override): schema
assembly must not perform blocking I/O.
A/B on the same machine, same dead endpoint: get_tool_definitions()
15.08s unpatched -> 1.89s patched, with browser_cdp/browser_dialog
still advertised (gate now keys off configuration, not reachability —
matching the documented lazy-supervisor contract in
_browser_dialog_check).
Adds a regression test asserting the browser_cdp check_fn never touches
the network.
The sidebar's flat session list lost its `x/<total>` chip in #72336, but
the project drill-in kept counting: WorkspaceHeader still rendered a
SidebarCount for every repo and every branch/worktree lane. Entering a
project put a number next to each label again.
Remove the header's count slot and both call sites, along with the now
dead repo total.
Home leads the overview above the active project and outside any drag
order, drills in to a flat chat list (it has no repo or worktree
structure), and overlays live sessions so a brand-new detached chat
appears instantly. Starting a chat from inside Home stays detached
instead of picking up the configured default project dir. Rename and
delete are hidden — there's no record behind the row.
Sessions with no cwd — or whose folder can't be promoted to a project (the
bare home dir, a deleted workspace, HERMES state) — were dropped from the
project tree entirely, so the grouped sidebar silently showed fewer chats
than flat Recents. Collect them into a synthetic `__no_project__` node at
the head of the list. It carries one lane purely to hold the rows, and is
omitted when empty so a project-less install stays blank.
The seam hairline sat at full strength on every split, so an empty
workspace read as a wireframe. Hold it at 0.1 and bring it up with the
grab band already on hover.
The active tab was defined by absence: the strip painted a rule and the
tab covered it, so inactive tabs stopped a pixel short to let it show.
Draw the state instead. The tab carries its own 2px --theme-primary
underline and the strip's rule goes away, which lets tabs run full height
and removes PANE_TAB_STRIP_LINE along with it.
The window chrome bracketed the workspace with a 1px rule top and bottom.
Both bars already paint the sidebar surface, so the rules divided one
continuous color rather than separating two.
The shell titlebar declared no background of its own, so it showed through\nto the wrapper's --ui-bg-chrome and read as a lighter band above the\nsession list. Use --ui-sidebar-surface-background, the token the sidebar\nalready paints with, so the two chrome surfaces meet on the hairline\ninstead of a color change.
The call-block decoration reads agent._use_prompt_caching / _cache_ttl /
_use_native_cache_layout directly; the redecoration helper wrapped each in
getattr with divergent defaults (e.g. or-'5m' vs verbatim _cache_ttl).
The flags are unconditionally initialized on AIAgent, so the defaults
served only test fixtures and would mask a real init bug as silent
cache-off. Align with the house style.
_peel_moa_guidance hand-implemented the inverse of moa_loop's
_attach_reference_guidance from a different module — a drifting separator
or shape would make the peel silently no-op and put the last cache
breakpoint on the turn-varying guidance block (the #72626 bug class).
Move the inverse into moa_loop.peel_reference_guidance directly adjacent
to the attach, keep a thin wrapper in conversation_loop, and pin the
contract with a round-trip test over all three attach shapes.
Also fix the empty-list residue: peeling a guidance-only content part now
drops the whole message (mirroring the appended-user-message shape)
instead of leaving an empty-content user turn behind.
guidance=None is a real prepared shape (all references failed / silent
degraded policy builds prepared_request without attaching guidance), and
the MoA facade sends prepared['messages'] — not api_kwargs['messages'].
Gating the rebase on 'and guidance' left the stale decoration in the
prepared object for the no-guidance MoA sub-path, so #72626 persisted
there. rebase_prepared_request already handles falsy guidance (copies
messages, skips the attach).
The static-prefix reconstruction pattern (build_system_prompt_parts ->
['stable'] -> startswith gate -> fail-open) existed in three copies:
session restore (conversation_loop), compression keep-prompt path
(conversation_compression), and the new failover redecoration helper.
Hoist it into agent/system_prompt.reconstruct_static_prefix and call it
from all three sites.
Also memoize failed rebuilds per stored prompt (_static_rebuild_failed_for):
the redecoration chokepoint runs at the top of every retry attempt, and a
persistent stable-tier mismatch (restored session whose SOUL.md/skills
changed since save) would otherwise re-run the full prompt build — SOUL.md,
context files, memory I/O — on every attempt of every API call for the
life of the session. A legitimately changed stored prompt retries once.
strip_anthropic_cache_control flattened ANY pure-text multi-part content
list with a separator-less join. Decoration only ever produces a single
text part or the 2-part [static, volatile] system split; organic
multi-part text (merged user turns, imported transcripts) got word-jammed
and parts carrying extra keys (citations) were silently dropped — on the
common no-failover path, since redecoration runs on every attempt.
Restrict the flatten to the exact decoration-produced shapes and make
marker removal copy-on-write on part dicts (the per-call message copy is
shallow, so parts alias the persistent history).
Add strip_anthropic_cache_control coverage and policy-change cases
(cache-off→on, on→off, native→envelope, MoA guidance outside marker)
that TestSyncFailoverPreservesCacheDecoration did not exercise.
try_activate_fallback refreshes the cache policy flags for the new
provider, but the retry loop reused the primary's decorated api_messages.
Cache-off→cache-on shipped zero breakpoints; cache-on→cache-off left
stale markers. Strip and re-render at each retry attempt (same chokepoint
as reasoning-echo reapply), peel/rebase MoA guidance so the last marker
stays off the turn-varying block, and rebuild the static system prefix
when caching becomes active mid-turn (#72626).
compress_context now runs on a daemon pool worker thread (via
run_compress_context_with_progress_timeout). The session id rotation
updates hermes_logging._session_context (a threading.local) on the
WORKER thread, not the caller thread. After the wrapper returns,
propagate self.session_id back to the caller's logging context so
subsequent log lines carry the rotated id (#34089).
Fixes CI failure in test_compression_logging_session_context.
Three code-reuse fixes applied during salvage:
1. Reuse _relative_time from hermes_cli/main.py instead of duplicating
the relative-time formatting logic in hermes_cli/status.py.
2. Extract _stamp_hygiene_compression_provenance helper in gateway/run.py
to deduplicate the two nearly-identical try/except blocks that stamp
compression timeout/abort provenance in the hygiene path.
3. Add ContextCompressor.record_timeout_failure() method and use it from
the in-agent compress_context timeout callback instead of re-implementing
the (60, 300, 900) cooldown ladder inline. The existing summary-LLM
exception handler already has this ladder — now both paths share one
method.
Three mechanisms to detect and notify when gateway sessions stall silently:
1. Mid-turn activity heartbeats stamped to SessionDB so hermes sessions list
and hermes status show progress during long turns without new message rows.
2. Stall watchdog: when a busy session has pending inbound and the shared
activity clock is idle past agent.session_stall_timeout (default 300),
log a WARNING and notify the user once to try /new. Notify-only; does
not kill the turn.
3. Compaction timeout: fenceless compress_context callers get a progress-aware
host budget (compression.context_timeout_seconds default 120 idle,
compression.context_total_ceiling_seconds default 600 ceiling). On timeout,
cancel via commit fence, skip compaction without dropping messages, and
continue the turn.
Closes#72016 (slices 1-3; slice 4 cumulative SSE stream-retry deadline
remains a follow-up).
Cherry-picked from PR #72424 by @fangliquanflq.
PR #72425 added getattr(agent, '_incremental_persistence_failed', False)
checks at the top of execute_tool_calls_{sequential,concurrent,segmented}.
A bare MagicMock auto-creates a truthy value for any attribute access,
so the interrupt-skip test's MagicMock agent short-circuited before
appending cancelled-tool messages — assert len(messages)==3 got 0.
Production is unaffected: run_conversation resets the flag to False
explicitly at turn start (conversation_loop.py:~1028).
- codex app-server sibling path: surface a WARNING (was silent debug) when
the projected-message flush fails — same bug class as the main fix, but
codex output has already streamed so fail-closed and agent_persisted=False
are both wrong here (#860/#42039 duplicate-write hazard); loud durability
gap logging instead.
- map session_persistence_failed in _format_turn_completion_explanation so
the user sees an actionable reason instead of 'The request failed:
unknown error' + explainer test.
- contributors/emails mapping for elco@thedaoist.gg (attribution CI).
Three helpers each re-derived part of the same decision: which backend
serves profile P, and does its REST path need a `?profile=` scope.
profileUsesPrimaryBackend answered the first half, pathWithGlobalRemoteProfile
answered the second, and ensureBackend re-checked globalRemoteActive() around
both. Splitting one table across three predicates is how the global-remote
case ended up registering reapable pool entries for a backend it never owned.
resolveProfileBackendRoute() states the four routes in one place and returns
the backend, the descriptor scope, and whether the path needs a query
parameter. The call sites read the answer instead of recomputing it.
One behavior change falls out: `hermes:api` now passes the primary profile
through, so the primary no longer sends itself a redundant `?profile=<self>`
on a global remote that already serves it.
Electron pre-installs its own uncaughtException listener and only warns on
unhandled rejections, so a main-process fault usually leaves the app running
with the reason on stderr — which nothing captures when the app is launched
from Finder or the Start menu. The fault never reaches desktop.log, so it is
absent from `hermes debug share` and the user can only describe symptoms.
Record both to desktop.log and flush synchronously, since a fault that does
prove fatal leaves no chance for the batched async flush. Five loadURL calls
were also unhandled, each able to leave a blank window with no explanation
anywhere the user can send us; they now name the surface that failed.
Co-authored-by: Rodrigo Fernandez <rod@nxtlevel.dev>
A pooled backend entry pointing at a remote host has no child process, so
the 'exit' handler that clears a dead local backend never fires. The
renderer's 60s keepalive touch also spares it from the idle reaper. Nothing
was left to retire the descriptor, so once the host went away the pool kept
serving it and every profile bound to that host stayed broken until restart.
Pooled remote descriptors now share the primary's liveness policy: probed on
the same revalidate tick, keyed per base URL, and dropped only after the
same consecutive-failure limit, so the next ensureBackend() rebuilds.
Co-authored-by: Rodrigo Fernandez <rod@nxtlevel.dev>
Keep non-primary profiles that inherit the app-global remote on the primary connection descriptor instead of creating processless pool entries that the idle reaper repeatedly removes.
Preserve per-profile remote overrides and local pooled backends, and cover the routing policy with behavioral tests.
Co-authored-by: Rodrigo Fernandez <rodrigo@nxtlevelsaas.com>
The change-detector test asserted 'hasattr(signal, "SIGPIPE")' in source.
PR #72677 replaced inline hasattr+signal.signal with _install_signal()
which does the same guard via getattr internally. Update the test to
accept either form.
Importing tui_gateway.entry from a worker thread raised
'ValueError: signal only works in main thread of the main interpreter'
because signal.signal() was called unconditionally at import time.
On the Desktop/WebSocket path, server._build() runs in a daemon thread and
does 'from tui_gateway.entry import ensure_mcp_discovery_started' as the
first import of entry (entry.main() is never run there), which crashed and
aborted MCP discovery startup — every session then lost its MCP servers
(e.g. Dart-mcp) with ClosedResourceError.
signal handlers are process-global, so installing them only when the module
is first imported in the main thread is sufficient; importing from a worker
thread becomes a safe no-op.
Fixes#72667
The stale-.pyc bug class (#6207, #60242, live WhatsApp report: gateway
ImportError 'cannot import name parse_model_flags_detailed' after /update)
has one shared shape: the checkout's .py files change while __pycache__
retains bytecode from the previous revision, and a later process trusts
the stale .pyc.
Update-time clears can never fully close this class: 'hermes update'
always executes the PRE-pull updater code, so hardening added to it takes
effect one update late — and a manual 'git pull' never runs the updater
at all.
Class fix:
- Launch-time guard in main(): compare the checkout fingerprint (cheap
file reads via _read_git_revision_fingerprint, no git subprocess)
against a .bytecode-fingerprint stamp; sweep __pycache__ once when they
diverge. Covers manual pulls, old updaters, ZIP restores — every entry
point (CLI, gateway service, desktop backend) passes through main().
- Record the stamp at all three update-time clear sites (git path pre-
install, git path post-install, ZIP path) so a normal update never
triggers a redundant launch sweep.
- Sibling site: 'hermes plugins update' + dashboard plugin update now
clear __pycache__ under the plugin dir after git pull (plugin trees
live outside the repo guard).
E2E validated: reproduced the stale-pyc shadowing (same-size same-mtime
source swap → old symbol wins in a fresh process), confirmed the launch
sweep restores the new symbol, no-ops when the checkout is unchanged,
and re-sweeps on the next pull. Sabotage-tested: reverting the sweep
fails 4 of the new tests.
Refresh update-sensitive modules before lazy backend refresh so an in-place git update does not keep using pre-pull module objects when newly pulled code imports fresh helpers.
Constraint: Issue #60242 reports post-update lazy backend refresh importing stale hermes_constants after a large Windows update.
Rejected: Only clearing __pycache__ earlier | the update process can still hold old modules in sys.modules.
Confidence: high
Scope-risk: narrow
Directive: Keep update-time lazy refresh guarded against in-process code skew after git pull.
Tested: ./.venv/bin/python -m pytest tests/hermes_cli/test_update_autostash.py::test_cmd_update_reloads_runtime_modules_before_lazy_refresh tests/hermes_cli/test_update_autostash.py::test_reload_updated_runtime_modules_restores_new_hermes_constants_symbol -q
Tested: ./.venv/bin/python -m pytest tests/hermes_cli/test_update_autostash.py -q
Tested: ./.venv/bin/python -m ruff check hermes_cli/main.py tests/hermes_cli/test_update_autostash.py
Tested: ./.venv/bin/python -m py_compile hermes_cli/main.py tests/hermes_cli/test_update_autostash.py
Tested: git diff --check
Not-tested: Windows v0.14.0 to v0.18.0 end-to-end update.
* fix(desktop): front workspace pane when navigating sidebar routes
Capabilities/Messaging/Artifacts (and other full-page workspace routes) rendered their content correctly on navigate(), but the workspace pane itself stayed behind an active session tile in the pane tree if one was focused. Clicking the same sidebar item again after switching to a session tile appeared to do nothing.
Session switches already call revealTreePane('workspace') + noteActiveTreeGroup(null) to front the pane (store/session-states.ts), but sidebar/keybind/command-palette navigation to full-page routes did not have an equivalent.
Add navigateToWorkspacePage() helper in routes.ts that wraps navigate() and fronts the workspace pane for non-overlay, non-chat views. Apply it at all 5 call sites that navigate to full-page workspace routes: selectSidebarItem, use-keybinds (nav.skills/messaging/artifacts), command palette 'go', Command Center's onNavigateRoute, and cold-start route restore.
Fixes#72602
* test(desktop): cover workspace pane reveal in selectSidebarItem
Adds a regression test for #72602: navigating to a sidebar route now calls navigate() and fronts the workspace pane (noteActiveTreeGroup(null) + revealTreePane('workspace')).
Mocks @/components/pane-shell/tree/store to assert the calls without depending on real pane-tree DOM state.
* fix(desktop): classify router targets by pathname, not the raw target
Every route classifier reasoned about the full navigation target, so a
query put them on the wrong branch: `/skills?tab=mcp` failed the reserved
path check and fell through to the session-id parser as the session
`skills?tab=mcp`, which made `appViewForPath` report Capabilities as a
chat. The command palette reaches Capabilities exclusively through those
targets, and Settings redirects old `/settings?tab=mcp` deep links there.
Strip the query and hash once, up front. Session ids are percent-encoded
by `sessionRoute`, so `?`/`#` can only ever start a query or a hash.
* fix(desktop): front the workspace pane from the router location
Capabilities, Messaging and Artifacts render inside the `workspace` pane,
so navigating to one has to bring that pane to the front of its group.
Nothing did. With the main zone parked on a session tile, the route and
the page content changed behind the tile and the click looked dead until
the app restarted. Session switches already front the pane in
`store/session-states.ts`; pages had no equivalent.
Front it from the router location, in the effect that already mirrors
`$workspaceIsPage`. One place decides, so every entry point inherits it:
sidebar, keybinds, palette, Command Center, contributed statusbar and
titlebar `to` targets, back/forward, and cold-start restore — which no
longer needs its own call.
`navigateToWorkspacePage` stays for the one case the location can't see:
hitting Capabilities while already on `/skills` leaves the location
untouched, so no effect fires and only an imperative reveal brings the
page back.
* test(desktop): cover the workspace-page reveal bug class
Both layers, and the classification underneath them: a page route fronts
the pane whether it carries a query or not, moving between two pages
fronts it again even though `$workspaceIsPage` never changes, contributed
routes count, and chat and overlay targets leave the tab alone.
---------
Co-authored-by: alelpoan <alelpoan@proton.me>
claude-fable-5 is a Mythos-class reasoning model (1M context, 128K output,
adaptive thinking per anthropic_adapter.py) but was missing from the
_REASONING_STALE_TIMEOUT_FLOORS table. Without a floor entry it got the
default 180s stale timeout (300s with context scaling), which is too short
for fable-5's thinking phase on large contexts.
Each stale kill bumped the cross-turn circuit breaker streak; after 5
consecutive kills _check_stale_giveup() fired immediately (elapsed: 0.00s),
aborting all calls with "Provider has been unresponsive for 5 consecutive
stale attempts." Users with 191K-token contexts hit this reliably.
Add ("claude-fable", 600) — deep-reasoning tier alongside o1/deepseek-r1/
nemotron-3-ultra. The claude-fable slug matches claude-fable-5 and future
variants via the existing right-anchor regex.
Both layers, and the classification underneath them: a page route fronts
the pane whether it carries a query or not, moving between two pages
fronts it again even though `$workspaceIsPage` never changes, contributed
routes count, and chat and overlay targets leave the tab alone.
