Composio eval traces showed Hermes wasting turns re-issuing identical tool
calls (same tool, same args, same result — 3x/4x in one run) and ending
turns by announcing an action it never took. Two conservative, config-gated
guards (agent.stall_guards, default true):
- Identical-call loop breaker: ToolCallGuardrailController.observe_identical_call
tracks the consecutive streak of (tool, canonical args, result-hash); on
the 3rd identical call a compact one-line notice is appended to that tool
RESULT at construction time (cache-safe — tool results are append-only).
Never blocks the call. Pollers (process, *_get_result, *_poll) are exempt
via STALL_GUARD_REPEATABLE_TOOLS. Streak resets on any different call,
changed result, or new turn. Observed on the raw result before the
tool-loop warning suffix so its changing count can't defeat matching.
- Said-continue-but-stopped recovery: trailing_continue_intent() detects a
short reply ENDING on an announced next action ('Let me now…', 'I will
now…', 'Next, I…'); the conversation loop feeds it into the EXISTING
intent-ack continuation path (same interim-assistant + user-nudge
mechanism, same codex_ack_continuations cap of 2), preserving message
alternation — no parallel recovery machinery.
Config: agent.stall_guards in DEFAULT_CONFIG; docs in configuration.md;
unit tests for streak/allowlist/reset/gate and detector pos/neg cases.
Composio-style MCP servers return un-paginated 22-47K-char payloads that
sail under the generic 100K per-result spillover threshold, bloating
context and ballooning per-turn reasoning time on long conversations.
Competitors cap harder (OpenCode/pi 50KB, Claude Code 30K, Codex ~10K
tokens). Three changes:
- mcp_* tools spill at a tighter 50K default (BudgetConfig.mcp_result_size,
config-overridable via tool_budget.mcp_result_size_chars; pinned and
per-tool overrides still win; capped by the context-scaled default).
- The persisted-output preview now teaches recovery: page the saved file
with read_file or process with execute_code instead of re-requesting the
same data from the remote API.
- Untrusted/MCP string results are scanned (bounded, first 64KB) for
provider-side elision markers ('...N more items', "has_more": true,
'saved to sandbox', data_preview) and get ONE cache-safe incompleteness
notice appended at result-construction time, before untrusted wrapping —
so the model stops treating provider-elided enumerations as complete.
- Hard 2M-char allocation cap in mcp_tool.py (text, error, and
structuredContent paths) so a pathological multi-MB server payload is
bounded before it propagates, while ordinary large results reach
spillover intact. Distilled from #56060/#56072/#56511 (issue #56059);
supersedes their 50K lossy truncation with spillover-friendly semantics.
Docs: configuration.md spillover-budget section + cli-config.yaml.example.
Co-authored-by: Stoltemberg <215755014+Stoltemberg@users.noreply.github.com>
Co-authored-by: AlexFucuson9 <295703459+AlexFucuson9@users.noreply.github.com>
Co-authored-by: Tranquil-Flow <66773372+Tranquil-Flow@users.noreply.github.com>
Unpinned zero-credential installs now pick Exa or Parallel by the
parity of the per-process random session id (stable within a process,
even split fleet-wide) instead of always favoring Parallel. An explicit
hermes tools selection (web.backend / per-capability keys) bypasses the
split entirely; the runner-up vendor stays in the walk as fallback.
Live E2E: 6 fresh processes split 3/3 between vendors, each performed
a real keyless search via its picked endpoint; explicit pin verified.
Un-fences OPENAI_MODEL_EXECUTION_GUIDANCE from the gpt/codex/grok substring
check and gives it its own injection gate, independent of
tool_use_enforcement, controlled by config.yaml `agent.execution_guidance`
(auto/true/false/list — same semantics as tool_use_enforcement). The "auto"
list (EXECUTION_GUIDANCE_MODELS) now also covers deepseek, kimi, qwen, glm,
minimax, mimo, and mistral.
Composio agentic-eval traces showed Hermes+DeepSeek/Kimi failing where
competitors passed: financial math done in prose, no read-back after
external writes, malformed identifiers "repaired", completeness claimed
despite count mismatches. The discipline block existed but those models
never received it.
The block is extended with compact clauses distilled from that analysis:
- external-write read-back (tool-call success is not task success; internal
file edits already confirmed by the tool are not re-verified)
- count reconciliation (declared totals/has_more are hard assertions)
- literal preservation (never normalize identifiers that fail a stated
format; lookup success does not validate a malformed token)
- retry-differently (empty/partial/suspiciously narrow results get a
broader retry before concluding)
- completion gated on verification (done = every named acceptance
criterion verified, never a plausible subset)
The todo tool description now encourages enumeration-as-checklist for
"all N items" tasks and gates completed status on verified work, never
intent.
Guidance is chosen once at session start keyed on model name, so the
system prompt stays byte-stable for the life of a conversation.
Supersedes/absorbs prior contributor proposals: #20588, #35087, #41874
(MiMo), #53847 (GLM tool-calls-as-text stall).
Co-authored-by: Mat-London <56627804+Mat-London@users.noreply.github.com>
Co-authored-by: intelac <8803887+intelac@users.noreply.github.com>
Co-authored-by: 6ylqq <51219463+6ylqq@users.noreply.github.com>
Co-authored-by: tauros1983 <267660491+tauros1983@users.noreply.github.com>
Exa and Parallel now each render as two picker rows in hermes tools —
'Free (keyless)' and 'Paid (API key)'. Selection persists to
web.provider_tier.<name>:
- free: always the anonymous public endpoint, even with a key set
- paid: always the keyed SDK path; missing key errors instead of
silently downgrading to the free tier (is_keyless_available also
returns False so the auto-fallback walk can't route there)
- unset: auto (key present -> paid, else keyless)
Mechanism: get_setup_schema() gains a 'variants' list the picker
flattens into sibling rows sharing one web_backend; selection writes
the tier via both _write_provider_config sites; active-row detection
matches the tier (auto mirrors use_keyless). Routing goes through a
single use_keyless() chokepoint shared by search+extract in both
providers.
Live E2E: tier=free with a fake key present searched keyless OK (a
keyed call would have 401'd); tier=paid without key errored naming
PARALLEL_API_KEY; picker rows verified for both vendors x both tiers.
With zero web credentials configured, web_search/web_extract previously
resolved to the nonfunctional firecrawl sentinel and errored. Now the
backend resolution walks a strictly-last keyless tier: Parallel's and
Exa's public anonymous MCP endpoints (the same free tiers opencode ships
as its default search path).
- plugins/web/keyless_mcp.py: minimal JSON-RPC tools/call client for
mcp.exa.ai + search.parallel.ai (SSE + plain JSON parsing, typed
errors, per-process random session id, no user identifiers)
- WebSearchProvider.is_keyless_available(): separate weaker tier that
never leaks into is_available(), so keyed setups are never pre-empted
- Exa/Parallel providers: route to keyless endpoints when their key is
absent; keyed SDK path unchanged
- registry + _get_backend(): keyless walk (parallel -> exa) strictly
after every keyed/importable candidate; check_web_api_key() lights
the tools up on zero-credential installs
- web.keyless_fallback config key (default true) to disable the tier
- docs: web-search.md + configuration.md
E2E-verified against both live endpoints from an isolated HERMES_HOME
(search + extract via the real dispatchers, disable-flag negative path).
Live incident 2026-08-17: the source checkout was parked on a stale feature
branch (claude-code-inspired/local-terminal-memory-limit, days behind main),
left there by earlier tooling. 'hermes update' autostashed, refreshed lazy
backends, synced skills, and printed '✓ Code updated!' / '✓ Update complete!'
while the checkout stayed on the stale branch with none of main's new code.
Two sessions burned time on 'the fix is missing' confusion.
- Parked-branch guard: auto-switch back to the update target ONLY when the
parked branch is clean and fully merged (git cherry origin/<target> shows
nothing unmerged); the checkout then STAYS on the target instead of being
re-parked. Otherwise: loud CODE UPDATE SKIPPED block naming the branch,
behind-count, and resolution commands; exit 1; branch untouched.
- The up-to-date (commit_count == 0) path no longer switches back to a
fully-merged parked branch either.
- Post-pull gate additionally refuses to print '✓ Code updated!' when HEAD
ends up attached to a non-target branch.
- Summary lines now carry the actual branch + HEAD short-sha:
'✓ Update complete! [main @ 30fcf9580]' — drift visible at a glance.
- New config toggle updates.auto_switch_parked_branch (default true).
- Real-git-fixture regression tests (init/clone/branch, no subprocess
mocks): clean+merged auto-switch, dirty skip, unmerged skip, cherry-picked
equivalence, config opt-out, unverifiable ref, on-main fast path,
up-to-date no-repark, summary branch/sha assertions.
Follow-ups on top of the salvaged CommandCode provider plugin (PR #32909):
- hermes_cli/config_defaults.py: COMMANDCODE_API_KEY setup-wizard entry
- hermes_cli/doctor.py: add key to the doctor env-var scan list
(health check comes free via the pluggable-profile loop)
- hermes_cli/dump.py: include commandcode in debug-dump api_keys
- docs: provider table row, fallback-provider table + supported lists
- tests: doctor dedicated-skip test now uses exact-name checks so
Bearer-authed Anthropic-COMPATIBLE gateways (CommandCode (Anthropic))
are allowed in the generic loop while native anthropic stays skipped
E2E verified with real imports: profile registration, aliases,
PROVIDER_REGISTRY auto-extension, bearer-auth host match
(positive + negative), live /models fetch (55 models).
The 'Configure auxiliary models' menu under 'hermes model' now includes a
Delegation entry so the delegate_task subagent model is discoverable and
configurable interactively, instead of requiring hand-edited
delegation.provider / delegation.model keys in config.yaml.
Delegation is not an auxiliary_client task — subagents are full child
agents resolved via tools/delegate_tool.py — so the picker entry writes to
the top-level delegation.* section rather than auxiliary.*. 'auto' (inherit
the parent agent) is persisted as empty strings, never the literal 'auto',
which delegate_tool would try to resolve as a provider name. 'Reset all to
auto' clears only the four delegation routing fields and preserves
non-routing settings like max_concurrent_children.
#87326 shipped the lean-compaction capability on the compressor; this adds
the config.yaml surface (compression.tail_mode: legacy|lean, default
legacy), DEFAULT_CONFIG entry, and docs on both the dev-guide compression
page and the user-guide configuration page, with zh-Hans parity.
Completes the duplicated-chrome class fix:
- Focus-in (CSI I) now routes through the same rate-limited full-redraw
recovery as Ctrl+L//redraw, clearing ghost prompt/composer copies after
Alt+Tab / tab switches (focus-regain variant reported on #60920, #25337)
- Document display.cli_rebuild_scrollback_on_redraw in configuration.md
- Register the new default in hermes_cli/config_defaults.py (moved from
the pre-refactor config.py location the salvaged commit targeted)
- Collapse the duplicated discord LoginFailure/PrivilegedIntentsRequired
classification (name-match + isinstance blocks repeated the same code/
message tuples) into a single _is() helper — one message per failure.
- Replace the user-facing HERMES_RECONNECT_ATTENTION_AFTER_SECONDS env var
with agent.reconnect_attention_after in config.yaml (default 7200, 0
disables), bridged internally like gateway_timeout. .env is for secrets.
- Use _float_env for robust parsing instead of bare int(os.getenv(...)).
- Document terminal classification + needs_attention escalation in
website/docs/user-guide/configuration.md.
Adds delegation.worktree_isolation (default: false). When enabled, each
delegate_task child gets its own git worktree branched from the repo's
current HEAD under <repo>/.worktrees/subagent-<id>, its terminal session
starts there, and its goal message carries the isolation contract
(work + commit in the worktree; parent reviews/merges the branch).
- tools/subagent_worktree.py: clean-room implementation from Muse Code's
documented --subagent-worktree-isolation behavior (create per-child
worktree, finalize/inspect after run, auto-prune clean no-commit
worktrees, keep anything holding work).
- tools/delegate_tool.py: config gate + per-child setup in
_run_single_child; result entries gain a "worktree" field (path,
branch, commits, dirty, pruned) only when isolation engaged — the
default-off wire shape is byte-identical.
- Git-only + local-terminal-backend-only; non-git dirs, remote backends,
or any worktree failure degrade silently to shared-workspace behavior.
- Tests: tests/tools/test_subagent_worktree.py (15 tests, real git
repos) + E2E through _run_single_child with a real repo verified
parent-checkout isolation, branch reviewability, prune, and
default-off shape pinning.
- Docs: delegation feature page section + configuration.md key.
OpenRouter's /v1/models entries advertise reasoning capability
(supported_parameters + reasoning.mandatory/supported_efforts). Use that
metadata as the primary gate in _supports_reasoning_extra_body instead of
the hand-maintained vendor-prefix allowlist, which went stale one vendor at
a time (nvidia/ missing -> #75386). Also clamp the requested effort to the
nearest LOWER catalog-supported level in the OpenRouter profile so ultra/max
against a high-capped route no longer 4xxes.
Cache-only on the hot path: capabilities parse for free out of the existing
fetch_openrouter_models() payload, a background warmer covers cold starts,
and unknown models/offline catalogs fall back to the static prefix list
unchanged.
Adds a pre_transcription transform hook (prompt/language/model mutable,
file_path read-only, last-writer-wins per the transform_* convention)
fired before any STT backend, threads prompt to faster-whisper
(initial_prompt) and OpenAI/Groq/Mistral/DeepInfra (prompt), adds an
optional stt.prompt config key on the same plumbing, and keeps the
no-hook dispatch path byte-identical. Fixes#64168.
Documents the new surface for users: a "Transcription prompt
(vocabulary hints)" subsection in the configuration guide (composition
order, per-provider support matrix, length contract, privacy warning),
a pre_transcription entry in the hooks reference, and the mirrored row
in the plugins hook table.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AG6LyYMvHC2o6HbVUozmVR
* fix: make verify_on_stop opt-in everywhere (default False, not auto)
The verify-on-stop nudge was already judged more noise than signal: the
v31 migration flips existing installs off, the v32 migration catches the
baked-in literal-true population, and the docs tell users to 'treat off
as the effective default and opt in explicitly'. But DEFAULT_CONFIG still
shipped the "auto" sentinel, so exactly one population kept getting the
nudges: fresh installs (and any config missing the key), where "auto"
resolves ON for CLI/TUI/desktop surfaces. Live symptom: repeated
'[System: You edited code ... run verification]' interruptions the user
never asked for and had to hunt down in source to disable.
- DEFAULT_CONFIG: agent.verify_on_stop "auto" -> False (opt-in).
- verify_on_stop_enabled(): missing/unrecognized value now falls back
OFF instead of surface-aware; explicit "auto" still selects the
legacy surface-aware behavior, explicit bools unchanged, and the
HERMES_VERIFY_ON_STOP env override is untouched.
- No migration needed: v31/v32 already normalized existing installs,
and this only changes the merged default for configs without the key.
- Docs updated; default-path E2E test now asserts OFF, plus a new
missing-value regression test. Also added the standard win32 skip
marker to the symlink-based temp-dir test (pre-existing Windows
failure, same class as tests/cron/test_cron_script.py).
* test: update config goldens — verify_on_stop=False is now stripped as default
With the DEFAULT_CONFIG flip to False, the migration-write invariant
(_persist_migration / save_config strip_defaults) no longer materialises
verify_on_stop: false to disk unless the user explicitly set the key:
- V20 floor fixture (agent: {} on disk): v31's write is stripped —
agent stays {} and load_config() supplies False at read time.
- V12 floor fixture (explicit verify_on_stop: true on disk): the key is
a user-set path, so the v32 flip stays materialised as false.
- Partial-write and _persist_migration regressions now assert the key is
absent from disk and (for the merge case) that the merged view still
resolves False.
Behavior verified with a one-shot migrate_config run against both
fixture shapes.
Two bugs reported on the docker terminal backend (desktop app, sandboxed
profiles with container_persistent: false):
1. A NEW chat's container inherited the PREVIOUS session's workspace,
bind-mounted rw at /workspace, because the mount source was the
process-global TERMINAL_CWD env var (written by the workspace picker,
outliving its session) and all sessions shared one 'default' container.
2. Every command failed with exit 126 because the desktop gateway recorded
the HOST launch directory as the session cwd, and each command was
prefixed with 'cd /Users/<user>/...' inside the container.
Fixes (class-wide, single owners):
- container_persistent: false + docker now keys containers PER SESSION:
fresh container per chat, removed at session close/idle. delegate_task
children share the parent's container via an explicit alias registry.
container_persistent: true keeps the documented ONE-long-lived-container
contract unchanged.
- _resolve_task_host_cwd() is the single owner of the cwd->/workspace mount
policy across all four env-creation sites; under isolation it refuses
process-global cwd sources and mounts only the session's own attached
workspace (tui_gateway now tags overrides with cwd_source).
- _resolve_command_cwd() gains the same host-path guard the env-creation
sites already had (#50636/#54447 sibling site): a recorded host cwd is
discarded on container backends instead of cd-ing every command into a
nonexistent path.
E2E-tested against real Docker: distinct containers per session, no stale
mount in a fresh session, no exit 126 from host cwd records, containers
removed at session teardown.
Record why the cache needs a third bound and what the pressure pass will and
will not shed, so an operator tuning agent.agent_cache knows which knob to
reach for. Adds the config keys to the session-lifecycle appendix and a user
guide section covering the "auto" cgroup-derived budget.
The docs promised 'frees ~370MB RAM' — measured behavior on macOS/CPU
is that ctranslate2's allocator keeps the freed pages (RSS doesn't
visibly shrink); the concrete win is VRAM release on CUDA hosts and
process-internal reuse on CPU. Say exactly that instead.
The local faster-whisper model singleton (_local_model) is loaded once
and never released — the 'base' model holds ~370 MB of RAM/VRAM for
the entire lifetime of the process, even when no voice messages arrive
for hours or days. On long-running gateway processes (especially with
local LLMs competing for the same GPU) this is wasteful.
Add a config-driven idle unload: after stt.local.unload_after_idle_seconds
(default 0 = never) of no transcription activity, a lightweight daemon
thread sets _local_model = None so the Python GC can reclaim the
ctranslate2 objects. The next voice message reloads the model
transparently (the existing lazy-load path handles it).
The watcher:
- Checks every 30s whether idle time exceeds the configured threshold
- Acquires _local_model_lock before unloading (prevents races with
concurrent transcriptions that are mid-load)
- Exits immediately if the model is already None (unloaded by another
path, e.g. the CUDA fallback eviction)
- Is restarted by each transcription with the current config value,
so changing stt.local.unload_after_idle_seconds in config.yaml takes
effect on the next voice message without a process restart
Default is 0 (never unload) — zero behavior change for existing users.
Recommended value for gateway processes: 300 (5 minutes).
15 tests: config resolution (garbage/negative/None fallbacks), unload
safety (already-None, lock acquisition), touch timestamp, watcher
lifecycle (unload after timeout, no unload within timeout, exits when
model already None, stopped on new start). Existing STT test suite
unchanged.
The review-fold commit added the input-duration gate but the user-facing
docs and config example still implied every cloud clip gets trimmed.
One-line additions to both.
Local faster-whisper gets Silero VAD (bf8004e3a) so silence never
reaches the model. Cloud providers got no such protection: the raw
file uploads untouched, so every second of silence in a voice note is
paid for twice — upload time and per-audio-minute billing — and cloud
Whisper hallucinates junk tokens on silent stretches exactly like
local Whisper did before the VAD hardening. A 13s voice note with two
long pauses is billed as 13s of audio to transcribe ~6s of speech.
Close the gap client-side: before uploading to a built-in cloud
provider (groq/openai/mistral/xai/elevenlabs/deepinfra), collapse long
pauses with ffmpeg's silenceremove filter, keeping
stt.cloud_trim_keep_ms (default 300) of every pause so word boundaries
and natural pacing survive. Uses ffmpeg, already a dependency of this
exact path via _transcode_audio_for_stt — no new dependency.
The trim is strictly best-effort — ALL of these upload the original
untouched, transcription never fails because of the trim:
- stt.cloud_trim_silence: false
- ffmpeg/ffprobe missing, trim failure, or timeout
- trimmed result ~empty (mostly-silence clip: the provider, not a
client-side dB heuristic, decides whether it contains speech)
- trim saves <10% (re-encoding for nothing)
Command-type and plugin providers are deliberately NOT trimmed: they
may wrap local CLIs that want the original bytes or run their own VAD.
E2E (real ffmpeg + faster-whisper): 13.2s voice note with 7s pause ->
6.2s upload (-53%); transcript of trimmed audio matches the original
on both utterances. Dense-speech and all-silence WAVs correctly fall
back to the original. 22 unit+E2E tests; STT/voice suite failures
identical to upstream/main baseline (all pre-existing).
The timeout table already lists the stale non-stream detector, but the
prose read as if it only guarded the interactive path. Spell out that it
bounds the inline cron / delegated-subagent calls too, and name the
accepted-then-silent failure mode it recovers from.
Follow-up to c0d974b19 (#79741). Three review findings against that commit,
none of which change the escalation behaviour it shipped.
1. The recovery decision lived inline in `_handle_message_with_agent`, a
~2000-line async method, so the only way to pin it was a test that read
`inspect.getsource(...)` and asserted on substrings. AGENTS.md bans reading
source in tests outright and names this exact situation: "if the logic lives
inline in a god-file (gateway/run.py) and extracting it feels disruptive:
that's the actual signal to do the extraction, not to regex around it."
Those tests were not merely stylistically wrong, they were actively harmful.
One asserted the substring `_new_tokens < _approx_tokens` was PRESENT -- so
it passed while the gate had the bug that substring represents, and had to be
edited when the gate was fixed. It failed on correct code and passed on
broken code, in one assertion.
Extracted `hygiene_compaction_recovered()` as a module-level pure predicate
and replaced the three source-reading tests with eight direct unit tests.
The extraction immediately earned itself: the new tests caught a `NameError`
(the predicate called `compression_made_progress` while the module bound it
under an alias) that a source-text assertion cannot see, because the symbol
is spelled correctly in the source and only fails at runtime.
2. The gate inferred "did the transcript actually get rewritten" from a numeric
side effect -- the degenerate "did not rotate or compact in place" path
(#21301) reuses the pre-compression counts -- when the booleans
`_hyg_rotated` / `_hyg_in_place` were already in scope and explicitly set
False on that path. The predicate now takes them directly, so a future edit
that re-estimates instead of reusing the old counts cannot silently defeat
the escalation.
3. `_record_hygiene_cooldown` passed no `error` to
`record_compression_failure_cooldown`, which writes `compression_failure_error`
unconditionally -- so a hygiene failure clobbered to NULL whatever reason the
in-conversation path had recorded, and readers then show the user "unknown
error" (agent/manual_compression_feedback.py, gateway/slash_commands.py). The
reason was already in hand at both call sites. Pre-existing from #74136 but
amplified by escalation: a blank reason on a 45-minute cooldown is far more
user-visible than on a 5-minute one.
Also: the ladder docstring described the compressor's absolute 60/300/900s
ladder while the constant is multipliers (1, 3, 9); the config docs still
described `hygiene_failure_cooldown_seconds` as a flat interval rather than the
first rung of a capped ladder; and `PersistentState.hygiene_failure_streak` now
documents that it is process-local by design -- keying on `session_key` is what
survives compaction rotation, which the persisted `compression_*_streak`
columns cannot express since they key on the rotating `session_id`. Making it
durable is a schema change, tracked on #79624 rather than smuggled in here.
Also replaces the file's hand-written `_Runner` stub with
`object.__new__(GatewayRunner)` (already the idiom elsewhere in the same file).
The stub reimplemented `_session_state` and `_peek_session_state`, so the tests
exercised copies that could drift from production; using the real class
immediately made one assertion stronger -- on a fresh runner `_sessions` does not
exist at all until something materialises it, so the reset provably did not even
create the map.
A review pass on this follow-up then caught that the CALL SITE was still
unbound: deleting the whole `if not _hyg_aborted: if
hygiene_compaction_recovered(...)` block left every ladder test green, because
the unit tests prove the predicate correct without proving it is wired in. The
merged commit had the same gap and its only cover was the banned source-reading
test. `test_session_hygiene_forces_in_place_compaction_with_bound_session_db`
now spies the reset on a genuine in-place compaction, so deleting the wiring
fails. Two earlier attempts at this test did NOT close the gap -- asserting on
streak VALUES passes either way, since the streak is 0 whether or not the gate
ran; only a positive spy assertion on a recovering run detects the deletion.
Same pass also corrected an overstatement: point (2) is hardening, not a live
bug. The degenerate path also sets `_new_count = _msg_count` and `_new_tokens =
_approx_tokens`, and `compression_made_progress(n, n, t, t)` is always False, so
the merged code already declined to reset there. A 200k-trial fuzz over the
reachable state space found zero behavioural disagreements between the merged
gate and this one. The guard's value is surviving a future edit that stops
reusing those counts.
Tests: 28 in tests/gateway/test_hygiene_failure_cooldown_ladder.py (8 new unit
tests for the predicate, 3 for reason forwarding, 3 source-reading tests
deleted). All 5 mutations caught -- including one that restores the hand-rolled
comparison and one that removes the rotated/in_place guard. Two mutations
initially SURVIVED and exposed vacuous tests of my own: the no-rewrite test used
counts the progress predicate already rejects, so it passed without binding the
guard at all; it now passes counts that read as progress on their own, proving
the guard is what rejects them. gateway hygiene + session-state + agent
compression-progress suites: 54 passed; ruff clean.
Refs #79624
The runtime footer (`/footer`) shows what model ran and how full the context
is, but not how long the turn took. On a messaging platform there is no
progress bar and no shell timer — a turn that took 4 seconds and one that took
four minutes produce visually identical replies. Users comparing models,
providers, or reasoning levels have no at-a-glance signal for the one
dimension they most often care about, and "was that slow or did I imagine it?"
is unanswerable after the fact.
Adds a `latency` field to the existing footer machinery, rendering the
wall-clock duration of the agent run: `<1s`, `22s`, `1m05s`.
`gateway/run.py` measures with `time.monotonic()` immediately around the
`self._run_agent(...)` await in `_handle_message_with_agent` — the same
function that already builds the footer, so the value is the user-perceived
turn duration (monotonic, so it is immune to wall-clock/NTP adjustment).
`latency` is deliberately NOT in `_DEFAULT_FIELDS`. It is opt-in via
`display.runtime_footer.fields`. Every existing footer — and every footer a
user has today without touching config — renders byte-identically.
This is enforced by tests, not just asserted:
- `test_latency_not_in_default_fields` pins the default tuple.
- `test_resolve_footer_config_default_fields_exclude_latency` pins what
config resolution produces for an untouched config.
- `test_default_footer_renders_byte_identically` pins five exact output
strings for default-config renders **while supplying `turn_seconds`** —
proving that even when the caller measures timing, a default-configured
footer does not show it.
- `test_default_build_footer_line_ignores_turn_seconds` asserts
`build_footer_line(...) == build_footer_line(..., turn_seconds=125.0)`
under default fields.
Adding `latency` to `_DEFAULT_FIELDS` fails 11 of these tests.
No new config surface (reuses `display.runtime_footer.fields`), no new env
vars, no new core tool, no new model-facing schema. One new module-private
helper (`_format_latency`), one new keyword argument threaded through the two
existing footer functions, and 3 lines in `gateway/run.py`.
`turn_seconds` defaults to `None` and the field is skipped when it is `None`
or negative, so any call site that does not measure timing keeps working
unchanged.
`tests/gateway/test_runtime_footer.py` (+185): `_format_latency` boundary
table (sub-second, rounding at 59.4/59.6, the `m{:02d}s` zero-pad, 60m), the
render/skip/opt-in matrix, field-order placement, `build_footer_line`
threading, and the byte-stability block above.
RED-proved by mutation — each of these breaks tests:
- `latency` added to `_DEFAULT_FIELDS` → 11 failures
- dropping the `turn_seconds is not None and >= 0` guard → 2 failures
- `{sec:02d}` → `{sec}` → 6 failures
- `build_footer_line` not threading `turn_seconds` → 1 failure
51 passed in `tests/gateway/test_runtime_footer.py`; 54 passed across the
footer blast radius. `ruff check` clean.
- website/docs/user-guide/configuration.md (en) and the zh-Hans
translation gain a 'Session Stall Watchdog' section: default 300,
0=disabled, notify-only semantics (never kills the turn — contrast
gateway_timeout), one notification per stall episode, and the exact
stall message text so it is greppable.
- cli-config.yaml.example: the two in-agent compression timeout keys
(compression.context_timeout_seconds /
compression.context_total_ceiling_seconds) are shown as commented
lines next to session_stall_timeout's example for discoverability.
The post-begin_commit() waiter previously called unbounded future.result(),
so the advertised compression.context_total_ceiling_seconds was silently
unenforced for commit-phase hangs. The commit still must complete (abandoning
an in-flight SessionDB mutation would diverge live messages from durable
state), but the wait is now bounded in increments against the remaining
ceiling: on ceiling breach the overrun is logged (WARNING escalating to
ERROR), surfaced once through the user-visible warning channel via the new
on_commit_overrun callback (wired to _emit_warning in run_agent.py), and the
host keeps waiting in bounded slices until the commit finishes.
Documented guarantee (config comment + docs, en/zh): summary phase bounded
by the ceiling; commit phase logged + surfaced if it exceeds it — never
silently hung, never abandoned mid-commit.
Test updated to assert the surfacing fires (previously accepted a silent
over-ceiling wait); adds coverage that a raising overrun callback cannot
break the commit wait.
Once begin_commit() wins, SessionDB mutation cannot be fence-cancelled;
document that context_total_ceiling_seconds covers the summary phase only
and pin the hang-wait contract in tests.
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.
Fixes#67278. Since config v12 (see #8776), custom_providers: list is
legacy — the migration converts it to the providers: dict and the
resolver reads the dict first (runtime_provider.py). Docs still taught
the legacy list as primary.
- integrations/providers.md: all 8 YAML examples converted to
providers: dict (field mapping code-verified: api, default_model,
transport), one consolidated legacy-format note
- configuring-models.md, configuration.md, credential-pools.md,
migrate-from-openclaw.md, provider-runtime.md, faq.md: examples and
prose flipped to dict-first; legacy list noted as still-read
- adding-providers.md untouched (sole mention is a literal test
filename)
Local faster-whisper called model.transcribe with bare {'beam_size': 5}:
no VAD, cross-window conditioning on, no confidence filtering. Pure
silence produced hallucinated tokens (E2E: 5s anullsrc WAV -> 'You',
no_speech_prob=0.705) and noisy clips could produce runs of junk, often
in other languages.
Three-layer class fix, one shared owner for every local-whisper call
site (build_local_transcribe_kwargs):
1. Silero VAD filter (bundled with faster-whisper) on by default —
silence never reaches the model. stt.local.vad: false restores the
raw behavior for music/ambient transcription.
stt.local.vad_min_silence_ms tunes chunk splitting (default 500).
2. condition_on_previous_text=False — one hallucinated token can no
longer seed a self-reinforcing run; negligible cost for
voice-note-length audio.
3. Segment confidence gate (_join_confident_segments): drop a segment
only when no_speech_prob > 0.6 AND avg_logprob < -1.0 (openai-whisper's
own heuristic shape; both must hit so quiet-but-real speech survives).
Config: stt.local.no_speech_prob_threshold / logprob_threshold.
The WHISPER_HALLUCINATIONS blocklist in voice_mode.py stays as
last-resort defense but should now almost never fire.
E2E (real faster-whisper 'base', CPU int8):
silence.wav before 'You' -> after ''
noise.wav before '' -> after ''
speech.wav before/after 'Hello World, this is a test of the
transcription system.' (unchanged)
Docs (EN + zh-Hans), DEFAULT_CONFIG, cli-config.yaml.example updated;
19 unit tests (kwargs contract, off-switch, confidence gate incl.
quiet-speech survival, _transcribe_local wiring), sabotage-verified.
Adds gpt-transcribe (OpenAI's new file-transcription model, $0.0045/min)
to the OpenAI STT provider:
- OPENAI_MODELS set: gpt-transcribe is recognized so provider
auto-correction keeps it on OpenAI and rejects it on Groq
- Language hint wiring: gpt-transcribe replaces the singular
'language' field with a 'languages' list; the API rejects the legacy
field, so the hint is sent via extra_body {languages: [..]}
- Config comment (DEFAULT_CONFIG), cli-config.yaml.example, desktop
settings enum, and docs (en + zh-Hans) updated
- Tests: model pass-through, languages-list hint shape, legacy singular
hint preserved for gpt-4o-transcribe, Groq auto-correction
gpt-live-transcribe (realtime WebSocket, $0.017/min) is NOT wired here:
the file-based STT pipeline has no realtime session path; it belongs in
a future realtime/voice-mode integration.
* Revert "fix(credits): remove the 'Grant spent · $X top-up left' notice"
This reverts commit 5dc6a14c14.
* fix(credits): reintroduce grant_spent behind an in-session crossing gate
The removed notice nagged because its condition is a steady state for
accounts living on top-up, the latch is per-session, and the cold-start seed
runs the policy at every session open — every session re-announced
'Grant spent · $X top-up left'. Reintroduce it gated so only a session that
WATCHES the grant run out announces:
- seen_grant_unspent crossing gate, mirroring seen_below_90: opens when the
session observes the grant meaningfully unspent (>= GRANT_UNSPENT_MIN_MICROS,
1 cent — portal-seeded states derive micros from float dollars and can carry
sub-cent residue where headers report exactly spent). Seeds never prime it.
- The gate guards only the show branch and is consumed by the announcement —
one announcement per crossing. Header flicker (used_fraction None and back)
clears the sticky line but cannot re-announce; a renewal re-opens the gate.
- new_credits_latch() centralizes the latch shape (agent build, lazy re-init,
and the policy test helper all build through it).
- Tests: steady-state-open stays silent (seed + policy seams), live crossing
announces once, flicker/renewal/residual cases locked; the rendering test
primes the gate and asserts its leg count so a gate regression cannot
silently shrink coverage.
Closes#55908. The CLI voice-mode beep amplitude is hardcoded at 0.3 inside
tools.voice_mode:play_beep(), which makes the record start/stop cues too
quiet on low-volume systems and headphones. Users couldn't adjust it
without editing source.
Move the literal into a configurable voice.beep_volume setting (clamped to
0.0-1.0, default 0.3 to preserve prior behaviour). The new
_get_beep_volume() helper reads via the same load_config() pattern used by
cli.py's _voice_beeps_enabled() and hermes_cli/voice.py's _beeps_enabled(),
keeps bools / out-of-range / non-numeric / NaN values safely on the default,
and falls back silently if config can't load so the audio cue never breaks
the voice loop on a degenerate config.yaml.
Covered by tests/tools/test_voice_mode.py:
- TestGetBeepVolume (12 cases: missing key, custom value, boundary 0.0/1.0,
out-of-range clamp, type coercion, bool guard, NaN guard, exception
guard, dict-typed voice section)
- TestPlayBeepVolumeWiring (guards against re-introducing a hardcoded 0.3
literal in play_beep)
Docs: website/docs/user-guide/configuration.md mentions the new key.
Other locale translations (zh-Hans etc.) intentionally untouched —
handled by the regular i18n sync pipeline as a separate change.
No change in default behaviour: existing users hear exactly the same beep.
The grant_spent notice fired for every subscription user with top-up
funds the moment their cap was reached and camped in the CLI/TUI status
bar and desktop toasts with no action to take — the account keeps
working off top-up. Remove it everywhere:
- agent/credits_tracker.py: drop grant_cond + the emit/clear block;
the dev fixture state now (correctly) produces no notice
- TUI: keep the turn-start clear of credits.grant_spent as back-compat
for older backends that still emit the key
- Desktop: drop the demo step and stale comment references
- Docs/config comments: remove grant-spent from credits_notices text
- Tests updated: policy/cold-start now assert the key never fires
Usage bands, depleted, and restored notices are unchanged; /usage still
reports the full balance breakdown.
An open chat window took a session-cap slot at session.create/resume time.
Every desktop tile paint and every background reconnect-resume opens one, so
on a websocket-flappy host they accumulated: five parked desktop tabs filled a
5-slot cap and locked the messaging gateway (which shares the cap) out for
fourteen minutes while running no agents at all.
A slot held that way is invisible everywhere. An unprompted draft has no DB row
and the sidebar filters it out with min_messages=1, so the only way to diagnose
it was reading runtime/active_sessions.json by hand.
Claim on the first turn instead, mirroring the lazy contract
_ensure_session_db_row already uses for the row itself. Capacity now means an
agent can run rather than that a window exists, and anything holding a slot is
something the user can see.
Also reclaim leases whose session skipped teardown. _prune_dead only fires when
the owning pid dies, and a dashboard/serve backend runs for days, so a leaked
lease was held until restart. The owning process reconciles against the leases
it still holds, which is exact and needs no heartbeat write on the turn path.
Whisper auto-detection frequently misidentifies short/accented clips,
which users experience as voice notes transcribed in the wrong language
(Teknium + CTO both hit this). The unified resolver from #73067 made a
global hint possible; this makes it the DEFAULT so stock installs stop
guessing. Non-English users set stt.language once; '' restores
auto-detect for multilingual use.
Deep-merge gives existing configs the new default automatically (no
_config_version bump needed); any explicit per-provider or global
language setting still wins.
Class-level fix for the 'STT transcribes the wrong language' issue family
(#55551, #50181 and siblings). Previously language handling was per-provider
chaos: local honoured stt.local.language, Groq/OpenAI/Mistral/DeepInfra sent
no language hint at all, xAI silently forced 'en', ElevenLabs used its own
language_code key, and there was no global setting.
- New _resolve_stt_language() helper: stt.<provider>.language >
stt.language (new global key) > HERMES_LOCAL_STT_LANGUAGE > auto-detect.
- Threaded through ALL providers: local, local_command, groq, openai,
mistral, xai, elevenlabs, deepinfra (shared OpenAI handler), command
providers, and plugin dispatch.
- xAI no longer forces English when nothing is configured (auto-detect).
- Mistral Voxtral now receives a language hint when configured.
- stt.groq.model is now honoured from config (previously env-only).
- DEFAULT_CONFIG gains stt.language, stt.groq, stt.xai, stt.mistral.language.
- Tests: tests/tools/test_stt_language_resolution.py (11 tests, sabotage-
verified) + full transcription suite green (236 passed).
Builds on cherry-picked contributor work from #19786 (@zombopanda),
#23161 (@materemias), #50684 (@BlackishGreen33).
Teknium