Accounts-tab cards derived from the unified provider_catalog() carry
status_fn=None and had no hardcoded branch in _resolve_provider_status,
so any future OAuth/account provider plugin rendered permanently
logged-out. Fall through to the canonical hermes_cli.auth.get_auth_status
slug dispatcher and adapt its shape, so membership AND status both
auto-extend with the hermes model universe.
Address review feedback on the keyVar test helper: it mocks one /api/env row
(an EnvVarInfo), so type it as such and mirror the sibling provider() factory's
base-plus-Partial-override shape instead of hardcoding positional args and
fabricated fields (description='X direct API', url=''). Route the WidgetAI test
through it too, removing the inline duplicate of the same object shape.
- API-keys tab: a SearchField filters provider cards by name / env-var key /
description, with a 'no providers match' empty state. Card order stays
priority-then-name (curated PROVIDER_GROUPS priority floats recommended
providers up; equal priority falls back to alphabetical).
- Accounts tab: 'Other providers' keep sortProviders order (priority, then
name) — unchanged.
Adds searchKeys/noKeysMatch i18n strings across all four locales. Vitest covers
priority/name ordering + live filtering + empty state.
Adds the end-to-end parity contract test: every CANONICAL_PROVIDERS entry (the
`hermes model` universe) must be configurable on a desktop Providers tab —
keys(/api/env) ∪ ids(/api/providers/oauth) ⊇ canonical. Asserted as an
invariant against the live endpoints so the GUI can never silently drift from
the CLI again.
Surfacing this contract caught Bedrock: it's aws_sdk (no api-key vars), so it
had no Keys card. /api/env now tags AWS_REGION/AWS_PROFILE to the bedrock
provider card. Anthropic is whitelisted as a legitimate dual-tab provider
(direct API key + subscription OAuth).
Also refreshes the _OAUTH_PROVIDER_CATALOG docstring to describe its new role
as the override base for _build_oauth_catalog().
buildProviderKeyGroups now groups provider env vars by the backend-supplied
provider/provider_label (from the unified catalog — the same identity hermes
model uses), falling back to the desktop PROVIDER_GROUPS prefix match only when
the backend gives no hint. A provider the backend tags now always renders its
own Keys card, even with no hand-maintained PROVIDER_GROUPS prefix row —
PROVIDER_GROUPS is demoted to a presentation overlay (priority/blurb/docs).
Adds provider/provider_label to EnvVarInfo. New vitest asserts a backend-tagged
provider with no prefix row still renders a card.
/api/providers/oauth now unions the explicit hand-tuned OAuth cards
(_OAUTH_PROVIDER_CATALOG — bespoke flow/status/cli, plus the api-key Anthropic
PKCE card and synthetic claude-code row) with every accounts-tab provider in
provider_catalog(). Any OAuth/external provider in the `hermes model` universe
now appears automatically, closing the drift where google-gemini-cli and
copilot-acp had no Accounts card despite being CLI-configurable.
Adds read-only status cards for google-gemini-cli (via existing
get_gemini_oauth_auth_status) and copilot-acp (managed-by-CLI, like claude-code).
DELETE handler routes through the same _build_oauth_catalog() builder.
Parity test asserts the Accounts tab offers every accounts-tab catalog provider
as an invariant.
The Keys tab now surfaces every keys-tab provider in provider_catalog() (the
`hermes model` universe), synthesizing a card even when the env var has no hand
entry in OPTIONAL_ENV_VARS. Closes the drift where openai-api, kilocode, novita,
tencent-tokenhub, and copilot were CLI-configurable but invisible in the desktop
Providers → API keys tab.
Each provider row now carries backend-derived provider/provider_label grouping
hints so the desktop can group by the same provider identity the CLI picker
uses. Hand OPTIONAL_ENV_VARS prose still wins where present (enrichment, not a
gate). Shared non-provider credentials (e.g. tool-category GITHUB_TOKEN) are
explicitly not hijacked into a provider card — Copilot uses its provider-owned
COPILOT_GITHUB_TOKEN.
Adds hermes_cli/provider_catalog.py, deriving one descriptor per provider from
the CANONICAL_PROVIDERS universe (what `hermes model` renders, auto-extended
from provider plugins), joined with auth/env from PROVIDER_REGISTRY and display
metadata from ProviderProfile (with canonical/env fallbacks for the four
profile-less providers and the many profiles with blank display/signup fields).
Each descriptor is tagged with the desktop tab it belongs on (keys vs accounts)
by auth_type. This is the single source of truth the desktop Providers tabs will
derive membership from, so they can no longer drift from the CLI picker.
Tests assert the parity contract (catalog == hermes model universe) and tab
routing as invariants, not snapshots.
The previous xAI auto-speech-tag tests asserted on the local
pause-only fallback and only passed because call_llm silently
returns None in the test environment. They gave zero coverage of
the new auxiliary-rewrite path added in the previous commit.
Add tests that:
- mock agent.auxiliary_client.call_llm and pin down the new contract
(auxiliary rewriter output wins over the local fallback)
- verify the system prompt lists every documented inline + wrapping
tag and uses BBCode-style [/tag] closing syntax
- cover markdown-fence stripping (with and without language hint)
- exercise the local fallback on rewriter exception, empty response,
None response, and missing-choices response
- confirm call_llm is NOT invoked when the input already has
explicit speech tags, or is empty / whitespace-only
- replace the end-to-end test that asserted on the silent-fallback
output with one that mocks the rewriter and asserts the
rewriter's tagged text is what reaches the xAI TTS API
Mirrors the existing Gemini TTS audio-tag rewrite path. When the input
has no explicit user/model speech tags, ask the configured auxiliary
model to insert a richer set of xAI-supported tags (laughs, sighs,
whispers, soft/loud, slow/fast, etc.) so voice-mode replies sound more
expressive. Falls back to the local conservative [pause]-only transform
on any auxiliary-model failure.
A plain /model <name> switch only lasted for the current session — every
new session reverted to the previously-configured model, so users had to
re-switch every time (e.g. glm-5.1 -> glm-5.2 on every launch).
Persist-by-default is now the behavior across all three /model surfaces
(CLI, gateway, TUI/dashboard), gated by a new config key
model.persist_switch_by_default (default true):
/model <name> switch model (persists to config.yaml)
/model <name> --session switch for this session only
/model <name> --global switch and persist (explicit, unchanged)
The effective persistence is resolved once via resolve_persist_behavior()
in hermes_cli/model_switch.py so --session opts out, --global opts in,
and the config-gated default applies otherwise. --global remains a valid
explicit no-op alias for the new default.
Commit 6724daa2c added refresh_interval=1.0 to keep the idle clock
ticking, but unconditional 1 Hz redraws in non-fullscreen prompt_toolkit
mode cause terminal emulators (Xshell, iTerm2, Windows Terminal) to
auto-scroll to the bottom on every tick — breaking scroll-up to read
history.
Drive it from display.cli_refresh_interval (0 = disabled, the default)
so users who want the ticking clock can opt in without affecting everyone.
Fixes: #48309
Related: 6724daa2c, 8972a151a
gui.log was registered in hermes_cli/logs.py::LOG_FILES (and surfaced by
`hermes logs gui`) but was never wired into `hermes debug share`. The share
report captured agent/errors/gateway/desktop tails plus full agent/gateway/
desktop logs — but nothing from gui.log, the surface the dashboard, TUI-over-
PTY bridge, and websocket layer (hermes_cli.web_server / pty_bridge /
tui_gateway) actually write to. A user reporting a dashboard or TUI bug shared
zero breadcrumbs from the broken surface.
Wire gui.log through all three share surfaces, matching the existing pattern:
- _capture_default_log_snapshots(): capture the gui snapshot (redacted like the rest)
- collect_debug_report(): add the gui.log summary tail block
- build_debug_share(): pull gui full_text, prepend dump header + redaction banner, add to the upload loop
- run_debug_share() --local branch: same, plus the local print block
- _PRIVACY_NOTICE: name gui.log in both bullets
Redaction is inherited for free — the gui snapshot goes through the same
_capture_log_snapshot(..., redact=redact) path, so secrets are scrubbed in
both the tail and full text (verified E2E: seeded key masked by default,
passes through under --no-redact, raw token never leaks).
Tests: seed gui.log in the fixture, add test_report_includes_gui_log, and bump
the upload-count tripwire 4->5 (test_share_uploads_five_pastes).
The built-in Piper provider (tts.provider: piper, Python piper-tts
package) already constructs piper.SynthesisConfig for the advanced
tuning knobs, but did not forward speaker_id from the user config.
This wires tts.piper.speaker_id through to SynthesisConfig.speaker_id
so multi-speaker ONNX models (e.g. libritts_r) can be addressed via
config without dropping to the command-provider path.
Changes:
- Add speaker_id to the has_advanced tuple so setting it triggers
SynthesisConfig construction (same gating as the other knobs).
- Pass speaker_id=speaker_id to SynthesisConfig. Defaults to 0
(Piper's own default; single-speaker models ignore the field).
- Tolerant parse: bad input (non-int strings, lists, dicts) is
dropped to 0 instead of raising. Booleans are rejected outright
(True/False would silently coerce to 1/0 and hide a config
mistake). Mirrors the same shape as the command-provider's
_resolve_command_tts_optional_number helper.
speaker_id is applied per-call via syn_config.speaker_id, so the
PiperVoice cache key is intentionally left as just (model, cuda) --
the same loaded model serves all speakers. Tests cover the
config knob, the tolerant parse, and the no-reload invariant.
sentence_silence is intentionally not added here: the Python
piper-tts SynthesisConfig does not expose that field (CLI-only).
/update calls dieWithCode(42) which tears down the gateway and
hard-exits the Node process — the same PTY-killing path that /exit
and /quit use. In the hosted dashboard chat there is no Python
update wrapper to catch exit code 42, and the PTY death bricks the
tab until a browser refresh.
Mirror the DASHBOARD_TUI_MODE guard that #48882 added for /exit and
/quit: refuse early with an explanatory message.
Address correctness gaps found in pre-PR review of the strict matcher:
- Profile selectors can appear on EITHER side of the `gateway` token
(`_apply_profile_override` strips `--profile`/`-p` from anywhere in argv
before argparse), so `hermes gateway --profile work run` and
`python -m hermes_cli.main gateway -p work run` are valid launches the
previous matcher wrongly rejected. Strip `--profile`/`-p`/`--profile=`/`-p=`
from anywhere before locating the subcommand.
- A profile literally named `gateway` (`hermes -p gateway gateway run`) made
the old token scan stop on the profile value; stripping the selector+value
first fixes it.
- Tokenize quote-aware with `shlex` so quoted Windows paths containing spaces
(`"C:\Program Files\Hermes\hermes-gateway.exe"`) are no longer split mid-path
and the dedicated-entrypoint match survives.
Without these, the matcher could MISS a real running gateway -> the opposite
failure (restart/status reporting "down" when up). Adds regression tests for
all three shapes.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The pyproject addopts pin `--timeout-method=signal` relies on signal.SIGALRM,
which doesn't exist on Windows. pytest-timeout raised AttributeError at timer
setup and aborted the entire run before any test executed, so the suite was
unrunnable on Windows by default. Override timeout_method to "thread" on
Windows in pytest_configure; POSIX keeps the more reliable signal method.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
`hermes gateway restart` on Windows could take the gateway offline with no
replacement. restart() was stop() -> sleep(1.0) -> start(), but the graceful
drain can run up to ~180s while the detached pythonw process stays alive. The
1s sleep let start() run against the still-draining old process; its
"already running" guard then no-opped, and when the old process finally exited
nothing relaunched it.
Two root causes, both fixed:
1. Loose PID detection. `_scan_gateway_pids` and the gateway.status helpers
used substring matches ("... gateway" in cmdline) for lifecycle decisions,
so they false-matched `gateway status`/`dashboard` siblings and unrelated
processes like `python -m tui_gateway`, plus stale gateway.pid records.
Add a shared strict matcher `looks_like_gateway_command_line()` in
gateway/status.py that requires the real `gateway run` subcommand (or the
dedicated entrypoints), and route `_looks_like_gateway_process`,
`_record_looks_like_gateway`, and `_scan_gateway_pids` through it.
2. restart() race. Wait until the gateway is authoritatively gone
(`get_running_pid()` + strict `_gateway_pids()`) before relaunch; force-kill
once if it lingers and raise rather than start a duplicate; verify the
relaunch produced a running gateway and raise loudly if not (no more
exit-0 silent outage).
Scoped to Windows; systemd/launchd restart paths are already drain-aware.
Adds tests/gateway/test_gateway_command_line_matcher.py.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Replaces the tautological test from the original PR (which asserted a
plain assignment it performed itself in the test body) with one that
exercises the actual contracts: _init_cached_agent_for_turn leaves
max_iterations untouched, and the per-turn IterationBudget rebuild
(turn_context.py) propagates a refreshed cap.
When a gateway agent is reused from cache, it retains the max_iterations
from its initial creation. If config.yaml agent.max_turns or HERMES_MAX_ITERATIONS
changed between turns, the cached agent's budget becomes stale.
Before reusing a cached agent, refresh agent.max_iterations from the
freshly-resolved value (read from env/config at line 14585).
Fixes partial issue from PR #48127: handles fresh agent creation + cached agent reuse.
The Discord fix (previous commit) handles dict-shaped clarify choices at the
Discord adapter only. The same dict-repr leak originates upstream at
tools/clarify_tool.py's str(c).strip() normalization — the single
platform-agnostic point both the CLI and every gateway adapter flow through.
When an LLM emits [{"description": "..."}] instead of bare strings, str(c)
produced {'description': '...'} which leaked onto the CLI panel
(cli.py:13048/13081), was returned verbatim as the user's answer
(cli.py:11945), and hit Telegram's numbered list too.
Add _flatten_choice (same label->description->text->title unwrap as the
Discord adapter, name/value excluded, keyless dicts dropped) and apply it at
the normalization line. Fixes CLI + Telegram + all platforms at the root;
the Discord smart-truncation now operates on already-clean text.
Adds johnjacobkenny to AUTHOR_MAP for the salvaged commit.
Two bugs surfaced from production usage in #37134:
1. Dict choices rendered as Python repr. LLMs sometimes emit
[{"description": "..."}] instead of bare strings; the old
str(c).strip() coercion turned the whole dict into
"{'description': '...'}" on the button label.
Fix: add a _flatten_choice helper that unwraps dicts against
the canonical LLM tool-call user-facing keys (label, description,
text, title) in that order. Dicts with none of those keys are
dropped. The "name" and "value" keys are deliberately NOT in the
priority list — they're Discord-component-shaped fields that
could appear in dicts that aren't meant to be choices (a
developer-error wiring that passes a Button-shaped object);
picking them would leak raw enum values or 4-char model
identifiers onto user-facing buttons.
2. Mid-word truncation on long button labels. The old
choice[:72] + "..." cut at position 72, mid-word. Worse, the
three-char ellipsis ate into the 80-char Discord label cap,
leaving only 75 chars of body.
Fix: budget-aware cut strategy with three tiers:
a. Last space in the trailing half of the budget (word boundary).
b. Last soft boundary (- , . )) in the trailing half — used
only when no word boundary exists.
c. Hard cut at the budget limit (last resort).
Use single U+2026 (…) to fit the cap. Cut AT soft boundaries
(inclusive) so the label ends on the boundary char rather than
on the alpha char that followed it.
Tests:
- test_unwraps_dict_choices_to_description: reproduces the
screenshot in #37134, asserts the Python repr is gone.
- test_unwrap_prefers_description_over_name_in_multi_key_dict:
regression guard for the name-key order in the unwrap list.
- test_unwrap_prefers_label_over_description: regression guard
for label winning over description.
- test_unwrap_does_not_pick_value_or_name_alone: regression
guard for the "name"/"value" fields being absent.
- test_truncates_long_choice_label: 200-char input, asserts
total <= 80 and U+2026.
- test_truncates_long_choice_label_breaks_on_word_boundary:
asserts the cut is on a space, not mid-word.
- test_truncates_long_no_space_choice_on_soft_boundary:
adversarial input where position 76 is mid-word alpha, asserts
the renderer falls back to a soft boundary.
Parity: telegram clarify suite (12 tests) still passes; the
helper is a Discord adapter local, not shared with the gateway.
Follow-up: gateway/platforms/telegram.py has the same str(c).strip()
pattern in its own send_clarify and will need a similar fix
(separate PR to keep this diff reviewable).
Fixes#37134
Unit-test `storedSessionIdForNotification`: runtime ids resolve to their
stored id, unknown ids and empty maps pass through unchanged, the right
stored id is picked among several sessions, and stored ids (map keys) are
never rewritten.
Native notifications (approval / sudo / secret / clarify) are tagged with
the gateway *runtime* session id — the key under which the session lives in
the gateway's in-memory `_sessions` map and the id every event carries
(`tui_gateway/server.py` `_emit(event, sid, ...)`). The chat route, however,
is keyed by the *stored* session id (`stored_session_id`), which is a
different value: a new chat gets its runtime id immediately but its stored id
only once the first turn persists.
`onFocusSession` navigated straight to `sessionRoute(<runtime id>)`, so
clicking a notification (e.g. an approval prompt) sent the route-resume path a
runtime id where it expects a stored id. `useRouteResume` then resumed it as a
stored session -> REST `/api/sessions/<runtime id>` 404 "session not found",
and the running session was navigated away, which the user experiences as the
session being destroyed.
Translate runtime -> stored before navigating via the existing
`runtimeIdByStoredSessionId` map (new `storedSessionIdForNotification`
helper), falling back to the id as-is when no mapping is known. The
Approve/Reject notification button path is untouched: `approval.respond` is
routed by the runtime id (`_sess()` -> `_sessions[session_id]`), so it must
keep carrying the runtime id.
_is_compression_ancestor walked parent links in a 100-hop Python loop
issuing two SELECTs per hop and hand-re-encoded the compression
continuation edge a fourth time. Collapse it into a single recursive CTE
that reuses the canonical _COMPRESSION_CHILD_SQL fragment (already shared
by _ephemeral_child_sql and set_session_archived), so the edge definition
lives in exactly one place. The UNION recursion also dedups visited nodes,
making it cycle-safe without the defensive hop cap. Behavior is unchanged
(all TestSessionTitleLineage + existing title-command tests pass).
Regression tests for renaming a compression continuation back to its base
title: single- and multi-level chains transfer the title off the ended
predecessor, while unrelated sessions and non-compression children (created
while the parent was live) still raise the uniqueness conflict.
When context compression rotates a session, the original is ended and the
continuation is auto-numbered (e.g. "name" -> "name #2"). The session list
projects the ended root behind its live tip, so the user never sees the
predecessor. But set_session_title's uniqueness check compared against ALL
sessions, so renaming the visible tip back to "name" dead-ended with
"Title 'name' is already in use by session <id the user can't find>".
When the conflicting title is held by a compression ancestor of the session
being renamed, transfer the title instead of raising: clear it from the
ended predecessor and apply it to the continuation. Uniqueness is preserved
(still exactly one session carries the title) and the parent-link lineage is
untouched, so resume-by-title and tip projection keep working. Genuine
conflicts with unrelated sessions, and with non-compression children
(delegate/branch), still raise as before.
The salvaged PR added the vitest devDep + config + a unit test but never
added a "test" script to web/package.json, so "npm run test" errored with
"Missing script: test" and the new suite was unrunnable. Add the script so
"npm run test" runs the suite as the PR body claimed (4/4 pass).
The dashboard's FastAPI server and a terminal CLI are separate processes
sharing one SQLite session DB; there is no inter-process push channel.
The Sessions page polled the 50 newest sessions every 5s for the
"overview" card but only re-fetched the paginated sessions list on page
change or delete, so a session started in a terminal never appeared in
the list until the user navigated.
Reuse the existing 5s overview poll as a change signal: when the head
session id changes, silently reload the current page (no loading
spinner flicker, no scroll/reset of expanded rows or bulk selection,
which are keyed by id). The detection logic is extracted into a pure
shouldRefreshSessions() helper with unit tests. Adds a minimal vitest
setup for web/ (test script + config).
TMUX is not forwarded over SSH, so a TUI launched on a remote host from
inside local tmux only sees TERM=tmux/tmux-256color with no TMUX var --
the cursor-drift bug still applies there. Extend supportsFastEchoTerminal()
to also fall back when TERM is tmux-flavored.
Deliberately scoped to tmux* only, NOT screen*: GNU screen sets the same
screen/screen-256color TERM and has no reported drift, so widening to
screen would disable the optimization for those users with no evidence of
a bug (matching the original PR's stated out-of-scope note).
Adds tests for tmux-flavored TERM (disabled) and screen/xterm TERM
(stays enabled) to guard against accidental widening.
When /goal (and other _PENDING_INPUT_COMMANDS: retry, queue, q, steer,
plan, undo) were typed in the TUI desktop app, slash.exec returned error
4018 instructing the frontend to fall back to command.dispatch. Some
clients failed that client-side fallback, leaving the command empty and
surfacing "empty command" — the user's typed text was silently dropped.
slash.exec now routes pending-input commands to command.dispatch
internally, eliminating the fragile client-side fallback hop. The
response is exactly what command.dispatch would have produced, so the
TUI client behaves identically once the round-trip succeeds.
Salvaged from #48944 — rebased onto current main. The original PR's
source change and test_goal_command.py update are correct, but it missed
the second test surface: tests/tui_gateway/test_protocol.py's
parametrized test_slash_exec_rejects_pending_input_commands still
asserted the old 4018 rejection for retry/queue/q/steer/plan, turning CI
red (5 failures). That test is rewritten here as a behavior contract:
slash.exec for a pending-input command must yield the same payload as a
direct command.dispatch call, and must no longer emit the old
"pending-input command" fallback rejection.
Co-authored-by: kyssta-exe <kyssta-exe@users.noreply.github.com>
`hermes backup` walked every file under HERMES_HOME, excluding only
hermes-agent / node_modules / __pycache__ / backups / checkpoints. Python
dependency trees (plugin and MCP-server venvs, site-packages) and pip/uv
tool caches that live under HERMES_HOME were swept in file-by-file,
ballooning a backup to hundreds of thousands of entries that crawl for
hours — the reported "backup stuck for days / 426543 files" symptom.
Add the canonical regeneratable-dir names (.venv, venv, site-packages,
.tox, .nox, .pytest_cache, .mypy_cache, .ruff_cache — mirroring
agent.skill_utils.EXCLUDED_SKILL_DIRS) plus .cache to the backup's
exclusion set, used by both run_backup and the pre-update/pre-migration
_write_full_zip_backup. .archive is intentionally left in so the curator's
restorable archived skills still get backed up.
Tests cover each new dir name (excluded at any depth), that .archive and
cache-resembling files are kept, and an integration check that a planted
venv/site-packages/cache is pruned from the actual backup zip while
skills/config survive.
