Agent Plugins v1 packages with 'streamable-http' mcp.json entries now load
through Hermes' existing URL-based MCP client instead of being reported and
skipped. The stdio-only limitation was the agreed follow-up slice from
PR #81196.
Boundary rules from the v1 spec (§7.2.1) are enforced:
- URL must be absolute http(s), no user information, no fragment; plain
HTTP only for localhost/loopback hosts.
- Configured package headers are never forwarded across a cross-origin
redirect: translation marks entries strict_redirect_headers, and the
redirect hook in the native runtime strips those headers (plus
Authorization) whenever a redirect leaves the original origin. On mcp <
1.24.0, where the client cannot hook redirects, such servers fail closed
with an actionable upgrade message.
- Legacy 'sse' entries remain reported and skipped.
The redirect hook is extracted into a testable module-level factory
(_make_redirect_header_stripper); default behavior for native config
servers is unchanged (Authorization-only stripping).
Document the actual transport split: legacy editable drafts by default, optional rich drafts, persistent rich final sends, and in-place rich final edits for edit-based streams.
New tests/skills/test_authoring_standards.py parametrizes every bundled +
optional SKILL.md (1148 checks) against the mechanically-verifiable subset
of the hardline standards:
- required frontmatter fields (name/description/version/author/license/
platforms) + tags
- frontmatter name == directory name
- description <= 60 chars, ends with period, no marketing words
- related_skills resolve in-repo
- no machine-local paths
- <= 100k chars
Grandfather dict for legacy debt ships EMPTY — all pre-existing violations
fixed in this PR:
- 13 frontmatter names canonicalized to their directory names (the install
identifier); all related_skills references updated (comfyui -> stable-
diffusion). Fixes the class behind PR #42788's report; also fixes
here.now's invalid dot-name.
- optional-skills/devops/cli -> inference-sh-cli (dir was the generic
'cli'; fm name was right) incl. docs pages (en + zh-Hans), catalog row,
sidebar entry.
- pytorch-fsdp: 157k generated 'Quick Reference' dump moved to
references/common-patterns.md; SKILL.md 159k -> 2.5k with a pointer.
- research-paper-writing: 31.7k Phase 5 drafting section moved to
references/phase5-paper-drafting.md; SKILL.md 103k -> 71k.
Docs regenerated with scope discipline.
Per review: upscaling should be the default behavior (like the original
flux-2-pro chain), not agent opt-in. Policy: every image model whose
native output is below ~2MP now sets upscale=True in its catalog —
users never silently get low-res images. Native hi-res models
(Seedream 5 Pro/Lite, Krea 2 Large) stay off to avoid paying to
upscale already-large output.
- FAL catalog: 16 models flipped to upscale=True (klein, z-image,
nano-banana pro/2/2-lite, gpt-image 1.5/2, ideogram v3/v4, recraft
v4/v4.1, qwen image/3, krea-2 medium on FAL, MAI 2.5 pro).
- Krea plugin: per-model upscale defaults (medium + medium-turbo ON at
1.5K native; large OFF at 2K native), precedence explicit kwarg >
image_gen.krea.upscale config > catalog default.
- The 'upscale' tool param remains as a per-call override in both
directions (false = fast draft, true = force on hi-res/edits).
- Video unchanged: opt-in only (default-on would double every video's
cost and latency).
- Sibling tests updated: routing/payload tests pass upscale=False where
the assertion targets the generation submit; catalog test now pins
the native-resolution policy instead of the flux-2-pro snapshot.
The generated-media surface previously had almost no upscaler coverage:
only fal-ai/flux-2-pro chained Clarity Upscaler (hardcoded catalog
default), every other image model returned ~1MP output with no high-res
path, and video had no upscaler at all. Krea's API treats the enhancer
as a standard second pass; this brings the same shape to Hermes.
- image_generate: new optional 'upscale' boolean in the tool schema.
Explicit true chains the backend upscaler on ANY model (including
edits); explicit false disables flux-2-pro's automatic default;
omitted keeps per-model catalog behavior. Response now reports
'upscaled' so the agent knows which resolution it got.
- FAL image path: explicit flag overrides the catalog 'upscale' default
(Clarity Upscaler, 2x). Failure falls back to the native image.
- Krea plugin: upscale=true chains Krea Enhance
(/generate/enhance/krea/enhance, 2x, prompt-guided) through the same
BYO/managed base URL + auth as generation, with a best-effort poll
loop that never fails a successful generation.
- video_generate: new optional 'upscale' boolean; FAL video plugin
chains ByteDance SeedVR2 (fal-ai/seedvr/upscale/video, 2x factor
mode). Providers without upscalers ignore the kwarg per the ABC
contract (documented in both ABCs).
Validation: targeted suites green (123 tests across 6 files, including
new coverage for override-wins/default-kept/failure-fallback on all
three paths); live E2E on direct FAL verified both chains end-to-end
(klein 9b + Clarity upscaled image; pixverse-v6 1s 360p + SeedVR2
upscaled video).
- description 210 -> 57 chars; author credits Ben Barclay (benbarclay) first
- optional-skills/creative/ (marketing vertical, narrowest audience of
the batch)
- dropped phantom 'image-generation-workflow' ref; visuals via the
image_generate tool
- honest handoff language: platforms without connectors end at approved
drafts marked handed-off, never claimed as published
- tests (10) incl. phantom-ref and honest-handoff guards
- docs regen scoped: per-skill page + one catalog row + one sidebar line
Skill polish:
- description 208 -> 57 chars; author credits Ben Barclay (benbarclay) first
- connector framing (google-workspace, obsidian, notion, email-inbox-triage)
- modern section order; boilerplate folded into step-local rules
Blueprint wiring (completes the batch's recipe integration):
- weekly-review blueprint loads weekly-review-planning; prompt follows the
skill's seven-section shape, drafts-only
- morning-brief blueprint loads google-workspace; prompt points at
references/daily-brief.md when connected
- important-mail blueprint loads email-inbox-triage
- blueprints index regenerated
Tests: 13 skill tests + two catalog invariants (every blueprint skills=
entry resolves to a real bundled skill; the four task blueprints are wired
to their procedure skills). 32 green across both files.
Opt-in via compression.codex_responses_native (default: false). When enabled,
gpt-5.6-family models on the direct OpenAI API (api.openai.com) or a ChatGPT
Codex subscription send context_management=[{type: compaction,
compact_threshold: N}] on Responses requests. OpenAI compacts server-side and
returns an encrypted compaction output item; Hermes captures it into the
existing codex_reasoning_items sidecar and replays it on later turns in place
of the pruned history — inheriting persistence, session replay, the
cross-issuer guard, and the encrypted-replay kill switch with zero new state.
Scope is deliberately hard-gated (agent/native_compaction.py, re-checked per
request): gpt-5.6 family only — gpt-5.1/5.2 fail server-side on the field
(HTTP 500 / stream stall, no structured rejection; live-verified) — and
direct OpenAI/Codex routes only; xAI, GitHub/Copilot, OpenRouter, relays,
and local servers never see the field.
Hermes' local compression stays armed as the fallback owner: the native
threshold is clamped ~8K tokens below the local trigger so the server
compacts first, and a structured provider rejection of context_management
disables native compaction for the session and retries without it
(one-shot guard in TurnRetryState).
Live-verified E2E on api.openai.com/gpt-5.6: server compaction fired at a
4K threshold, checkpoints captured and replayed, recall preserved across
3 turns; gpt-5.1 with the flag enabled stays clean (field never sent).
Direction credit: PR #76950 by @laryhorb explored native Responses
compaction; this is a minimal reimplementation on current main.
The bundled docx, xlsx, powerpoint, and pdf skills were adapted from
Anthropic's document skills and carried their proprietary LICENSE.txt
(no derivatives, no redistribution). Flagged as critical license
findings by the SkillEvaluator Tier 1 scan of our skill tree.
This replaces all four with clean-room rewrites:
- Authored from scratch against library knowledge only (python-docx,
openpyxl, python-pptx, pypdf/reportlab/pdfplumber — all MIT/BSD) by
isolated subagents given functional specs, with an explicit
prohibition on reading the prior skill content or anthropics/skills;
session transcripts retained as provenance evidence.
- MIT licensed (LICENSE file per skill), author: Nous Research.
- Each skill: SKILL.md to house standards + argparse helper scripts
with UTF-8-explicit I/O + its own e2e pytest suite (fixtures built
on the fly, non-ASCII round-trips run under LC_ALL=C).
- All four pass SkillEvaluator Tier 1 pii+unicode+lint 3/3.
tests/skills/test_office_document_skills.py rewritten against the new
contracts: MIT/no-Anthropic-text invariants, scripts documented in
SKILL.md, argparse CLI shape, and a no-locale-default-open() check
(which caught and fixed a real gap: pdfplumber text reads are fine,
but the invariant scan now guards every future script).
Docs pages regenerated for the four skills (scoped; unrelated
generator drift excluded).
Honest capability deltas vs the old versions are documented per
SKILL.md (e.g. tracked-changes accept/reject and OOXML XSD validation
are not reimplemented; form flattening limits stated).
Salvage of PR #17973 by @TKCen (Sebastian Hänisch), re-implemented on
current main to preserve speed/instructions/provider params,
prepare_spoken_text normalization, OPUS_VOICE_PLATFORMS, is_write_denied
path security, microsecond timestamps, and the streaming-TTS gate.
- Split long TTS text into provider-safe chunks instead of truncating
- Pack generated audio against platform upload limits (Discord 10MB,
Telegram 50MB, configurable via tts.delivery_profiles)
- Combine chunks with ffmpeg (OGG/Opus re-encoded, MP3 stream-copied)
- Multi-file delivery when combination fails or would exceed limits
- Remove hard [:4000] truncation from all callers (cli.py, voice.py,
gateway/run.py, gateway/platforms/base.py)
- Gemini TTS raises ValueError instead of silently truncating when
composed prompt exceeds the provider limit
Simplify-code fixes: removed dead all_touched_paths set, added
try/finally for scratch file cleanup on exception, clean error response
on chunk failure instead of leaking stale file_path.
The published tarball ships lib/binding/napi-9-darwin-unknown-arm64 on every
platform, so a real Windows host has both it and the downloaded win32 binding
— the classify-everything gate threw on the darwin dir and killed every
Windows pack. Stage only bindings naming the target platform (classify still
rejects impostors), stop copyGlobByExt from recursing into lib/binding, and
add a version tripwire so a get-windows bump fails the build until the
lib/windows.js rewrite is re-verified.
Also from review: the renderer answers window.read.respond with empty text
when the IPC invoke rejects (older shell / main-side throw) instead of
stalling the tool's 30s timeout; the tool schema discloses that sibling
Hermes windows are skipped; docs gain read_window_below in both references.
The brief is single-connector (every command comes from google-workspace),
so it ships as references/daily-brief.md with a pointer + load trigger in
SKILL.md — progressive disclosure instead of a new skill-index entry.
Contributor's procedure preserved (half-open day windows, mail-to-meeting
linking with fuzzy-match discipline, 7-section brief, bounded actions);
credit noted in the reference header. Tests (8) guard the wiring and
disciplines. Version 1.1.0 -> 1.2.0.
The coverage warning listed bare page ranges, which tells the agent
WHERE the gaps are but not WHAT they contain — its only options were
guessing or OCRing everything. Each gap is now labeled with the last
text extracted before it (usually a section divider page), so the agent
can decide which gaps it actually needs and render/OCR only those.
Gap list capped at 20 entries with a summary line for pathological
alternating documents.
Rewrote from a sibling-skill routing table into a skill that carries its
own procedure, and folded in generalized rules from maintainer practice:
- full-thread reads (gh issue view --comments; newest comment = live state)
- duplicate-PR sweep (issue number + keyword variants) before any code
- design-intent check via git log -p -S alongside premise reproduction
- fix the class: sweep sibling call sites into the same PR
- sabotage run: prove the regression test fails without the fix
- open the PR immediately (PR dispatches CI; CI latency is the long pole)
- close the loop: comment the issue with the PR link
Also: description 205 -> 59 chars, author credits Ben Barclay first,
modern section order, boilerplate trimmed, tests (10) incl. a
router-pattern guard, scoped docs regen.
The non-streaming /v1/responses path built function_call and
function_call_output output items with no status field (and no item id),
while the SSE streaming path correctly emits status in_progress ->
completed. Spec-strict OpenAI clients reading the non-streaming output
array could interpret the status-less function_call items as pending
calls the CLIENT must execute — but these tools were already executed
server-side by the Hermes agent and are replayed for structured tool UI
only. Reported by a community user whose GPT-5.6 client concluded 'a
server should not tell an OpenAI client to execute a tool the server
already executed itself'.
- _extract_output_items now stamps status: completed and spec-shaped
item ids (fc_/fco_) on replayed items, matching the streaming path
- test updated to pin status + id shape
- docs example updated + explicit note that output tool calls are
replayed, never pending
- description 219 -> 58 chars
- author credits Ben Barclay (benbarclay) first
- modern section order; trimmed template safety boilerplate into
step-local rules and a skill-specific verification checklist
- tests at tests/skills/test_email_inbox_triage_skill.py (9 passing)
- docs regen scoped: per-skill page + one catalog row + one sidebar line
--resume latest resolves the most recent session through the same
workspace-scoped MRU lookup as -c (TUI source first under --tui, with
classic-CLI fallback). --in DIR chdirs before session resolution so the
lookup keys off DIR's workspace, and pins the session there by skipping
the recorded-cwd restore.
Requested by @Jeff9James: hermes --tui --resume latest --in ./dir
Fleet audit showed these task skills are commonly needed across users;
shipping bundled per Teknium's direction. Docs and tests follow the
bundled paths.
Adds a per-server `trust: full|untrusted` config key
(mcp_servers.<name>.trust). On an untrusted server, every write-capable
tool call — any tool whose discovery-time annotations do not carry
readOnlyHint=True — routes through the existing approval surface
(tools.approval.request_elicitation_consent, same lazy-import +
surface-routing pattern the MCP elicitation handler uses) before the RPC
fires. Denied/cancelled/errored approvals fail closed: the RPC never
runs, including the lazy first-use server spawn.
Design points:
- Classification happens at CALL TIME from metadata captured at
DISCOVERY (_record_tool_trust_metadata in _register_server_tools and
the lazy cache-registration path). No toolset/schema mutation, so the
toolset stays byte-stable and prompt caching is preserved.
- readOnlyHint is a server-supplied HINT: on an untrusted server a lying
server can at most skip approval for tools it claims read-only — it
can never widen access. Trust tiering itself is operator config.
- Missing/malformed annotations => write-capable (fail closed).
- Unrecognized trust values => untrusted (fail closed); missing key =>
full (backward compatible, documented in mcp-config-reference).
- The schema cache now persists readOnlyHint so lazy-registered servers
gate identically on next startup without spawning.
Tests: tests/tools/test_mcp_trust_gating.py (11 tests, TDD red->green):
approval invoked + accept proceeds, deny/cancel blocks RPC, readOnlyHint
=true skips gate, trusted/unconfigured servers skip gate, explicit
readOnlyHint=false gated, approval exception fails closed, trust
normalization, discovery-time capture (SDK objects and cached dicts).
Ported from: cloudflare-os classifyTool() (Apache-2.0), corroborated by
Claude Cowork (idea-level).
Two small config-gated features:
1. Kanban orphaned-card reconciliation (kanban.reconcile_orphans, default
true, config.yaml): a running card with broken claim bookkeeping
(claim_lock or claim_expires NULL — crash mid-claim, manual SQL, DB
restore) is invisible to all existing recovery paths
(release_stale_claims requires claim_expires NOT NULL,
detect_crashed_workers requires host-local lock + pid,
detect_stale_running is config-disabled by default) and shows Running
forever. New reconcile_orphaned_running() pass in kanban_db.py runs
each dispatch_once tick: requeues orphans to ready with an explanatory
comment, closes any leaked run, emits a 'reconciled' event, and defers
when the recorded PID is still alive on this host (never requeue
beside a live worker). Surfaced via DispatchResult.reconciled_orphans.
2. Per-server MCP identity header (mcp_servers.<name>.identity_header,
config.yaml): optional {name, value_from: static|profile, value}
mapping; the header is attached to that server's HTTP/SSE transport
requests. 'static' sends the config value; 'profile' resolves the
active Hermes profile name once at connect time (no per-call
mutation). Explicit per-server headers of the same name (any casing)
win. Invalid blocks warn-and-ignore; stdio servers warn-and-ignore.
Tests: tests/gateway/test_kanban_reconcile_orphans.py (9),
tests/tools/test_mcp_identity_header.py (13), all written first (RED)
then implemented (GREEN). No new HERMES_* env vars.
Inspired by: openai/symphony tracker reconciliation (Apache-2.0) +
Poke per-user MCP identity (idea-level).
Validate a job's configuration BEFORE any agent machinery is constructed:
- missing provider API key (AuthError from a read-only
resolve_runtime_provider probe; skipped when a fallback_providers chain
is configured, since auth-fallback may rescue the run)
- attached skill not ready (skill_view readiness_status=setup_needed —
missing required env vars / commands / credential files)
- delivery platform unknown or unconnected (deliver=local/origin/all are
never checked; gateway-config load failures fail open)
On a failing check run_job returns a [blocked_config]-marked error without
constructing AIAgent/MCP/etc, so a misconfigured job never burns an LLM
call. run_one_job records last_status='blocked_config' and delivers the
alert exactly ONCE across ticks (persisted preflight_alerted bit — the
alert-once shape from the #73506 dead-pin auto-pause); the next healthy
run clears the marker so a future break re-alerts. Every preflight check
fails open: only an affirmative misconfiguration verdict blocks.
Config: cron.preflight (default true); `cron.preflight: false` restores
the old fail-during-run behavior. Documented in the cron user guide and
config defaults.
mark_job_run gains an optional status= override (unblocked call shape
unchanged) and drops preflight_alerted on any successful run.
Tests: tests/cron/test_preflight_config.py (blocked_config + no agent +
single alert across two ticks, healthy job unaffected, recovery clears
dedup, fallback-chain rescue, opt-out restores old behavior, skill
readiness miss, unknown delivery platform, deliver=local never loads
gateway config). Full tests/cron/ + cronjob tool suite green (525 tests).
Ported from: paperclipai/paperclip execution-semantics §5 (MIT);
in-repo precedent: #27948, #73506
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).
External hosts speak this protocol directly, so the parameter that
rewrites a session's stored transcript should not be folklore. Document
what each truncation field means, that an ordinal without
confirm_truncate is refused, and that a client must never hold the
ordinal in state across ordinary submits.
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.
- matrix/dingtalk: extract deps-only installers (ensure_matrix_deps,
ensure_dingtalk_deps) and register THOSE as ensure_deps_fn — the prior
check_*_requirements combined credential env checks with the install,
so a platform configured via PlatformConfig.extra (which is_connected
accepts) would pass enablement, reach create_adapter(), and have the
'installer' veto on env-var grounds before installing anything —
re-creating the #79812 deadlock for extra-configured setups. The
combined deps+credentials functions remain for setup/status callers.
- matrix/feishu passive probes: use the existing lazy_deps.is_available()
instead of hand-rolling 'not feature_missing(...)' (reuse finding).
- teams: module docstring no longer recommends bare system pip (the
PEP 668 trap purged everywhere else); docs troubleshooting row updated
to match the new hint text.
- wecom_callback: drop dead 'global ET, DEFUSEDXML_AVAILABLE'
(ensure_and_bind mutates the module dict directly; nothing assigns).
- tests: parametrized wiring contract for all 8 lazy-installable
platforms — ensure_deps_fn present and distinct from check_fn
(behavior contract, not identity snapshot, so renames don't churn it).
- gateway/config.py: rewrite the stale enablement-pass header comment that
still described check_fn as 'the single source of truth for are-my-env-
vars-set' / 'lazy-installs it' — both false under the new contract.
- teams: check_requirements docstring wrongly claimed credential checks
(body checks only SDK/aiohttp presence); derive install_hint from the
canonical LAZY_DEPS pins + sys.executable instead of hardcoding
'~/.hermes/hermes-agent/venv/bin/pip' and version pins (wrong under
HERMES_HOME overrides / profile installs; pins go stale on CVE bumps);
connect() fatal-error hints now point at the venv pip instead of bare
system pip (the PEP 668 trap the docs warn about).
- teams docs: drop exact version pins from the two manual-install commands
(LAZY_DEPS is the source of truth; unpinned installs still work and the
text can't go stale).
- hermes_cli/status.py: per-entry exception guard around check_fn so one
raising probe can't abort the listing of all remaining plugin platforms
(aligns with the other three call sites).
- tests: rename test_register_check_fn_is_active_lazy_installer ->
test_register_splits_passive_probe_from_active_installer (name said the
opposite of what it verifies).
PlatformEntry.check_fn served three contradictory roles: adapter-creation
gate, config auto-enablement gate, and status display. Plugins had to pick
one function for all three:
- Active installer as check_fn (discord/slack/telegram/matrix/dingtalk/
feishu): every status display could pip-install SDKs as a side effect
(the desktop 94% boot-loop class).
- Passive probe as check_fn (teams, wecom_callback): create_adapter()
returned None before connect() could lazy-install, so the SDK never
installed (#79812 deadlock; wecom_callback's platform.wecom_callback
LAZY_DEPS entry was dead code).
The split makes both call sites correct by construction:
- check_fn is now contractually PASSIVE (probe only, never installs).
- New optional PlatformEntry.ensure_deps_fn is the ACTIVE installer;
create_adapter() runs it exactly when check_fn is False — the platform
is enabled+configured and the gateway is about to connect it.
- Config enablement keeps a configured platform whose deps are missing
but installable; the install itself is deferred to create_adapter().
- Status surfaces (_platform_status, hermes status) read only the
passive probe and can never trigger pip.
Migrated all lazy-installable platform plugins to the split; platforms
with no optional deps (irc/ntfy/buzz/simplex/line/a2a/...) are unchanged
— no ensure_deps_fn means a False check_fn stays a hard block.
wecom_callback gains a working installer for the first time.
Builds on @xxxigm's #79812 (both commits cherry-picked with authorship
preserved), reworking the check_fn swap into the two-field split so the
Teams fix doesn't reintroduce install-on-status.
Covers the behavior shipped in #80807 (background dispatch for
cronjob action='run') and #80838 (per-run '## Run Context' prompt,
gateway-loop delivery): immediate return with handle, completion
re-entering the conversation, in-flight dedupe, transient context
injection with prompt scanning, and the sync fallbacks.
Inspired by virgiliojr94/book-to-skill (MIT): /learn now picks the skill
shape by the source. Workflows and small sources still get one tight
SKILL.md; books, paper stacks, specs, and large doc corpora get a
knowledge-base layout — a lean always-loaded SKILL.md index plus one
distilled file per chapter/topic under references/, loaded on demand via
skill_view so query cost stays proportional to the answer.
- agent/learn_prompt.py: new _KNOWLEDGE_SKILL_STANDARDS block (index +
per-chapter references/, structure-not-summary distillation, never
reproduce source passages, fold-in instead of duplicating) and a
_SOURCE_HYGIENE block pinning extracted source text as data and
dropping invisible/bidi Unicode (Trojan Source class). Clarified that
the ~200-line cap and hub-skill ban apply to SKILL.md itself, not a
knowledge skill's own references/ files.
- tests: contracts for the knowledge-base layout, the three embedded
standards blocks, and the source-hygiene coverage.
- docs: skills.md documents the knowledge-base shape.
The in-repo skill-authoring skill taught the validator's ceilings (1024-char
descriptions, 'Use when ...' phrasing) instead of the repo's review standards,
so agents following it produced skills that fail review: 240+ char
descriptions, author 'Hermes Agent' with no human credit, no bundled-vs-
optional decision, dangling related_skills, no platforms audit, no tests, no
docs regen, and machine-local /home/bb/... paths baked into prose.
Rewritten to teach the hardline standards from AGENTS.md:
- description <= 60 chars, one sentence, ends with period
- author credits the human contributor first
- bundled vs optional tier decision (5+ sessions/month bar; default optional)
- no router/index/hub skills
- platforms: audited against actual scripts, POSIX-signal table
- related_skills must resolve in-repo
- Hermes-tool framing instead of raw shell prose
- tests at tests/skills/ + docs regen with scope discipline
- removed machine-local paths; validator limits marked as NOT the standard
The six GUI tools moved out of `terminal` into their own toolset; the tables
still described them as check_fn-gated members of it, and as available to every
hermes-* platform bundle.
Per the 'when in doubt, optional' rule — niche prediction-market data
skill that sees no regular use; belongs alongside stocks in the finance
optional category rather than the default bundle.
Install via: hermes skills install official/finance/polymarket
The turn finalizer already hands back steer text that queued after the
final tool batch — result["pending_steer"], with the comment "hand it
back to the caller so it can be delivered as the next user turn instead
of being silently lost." Every interactive surface honors that contract
(cli.py, gateway/run.py, tui_gateway/server.py all requeue it). The
delegation layer doesn't: _run_single_child never reads it, so a steer
queued into a delegated child that finishes first vanishes with no trace
in the completion entry. There is also no sanctioned sender: the registry
has interrupt_subagent() but no redirection-side mirror, and session.steer
cannot reach children (lazy watch sessions have agent=None, so it 4010s).
Complete the contract for delegated children — both halves:
- steer_subagent(subagent_id, text): redirection-side mirror of
interrupt_subagent(). Resolves the live child in _active_subagents and
queues text via AIAgent.steer(). True means queued, not delivered.
- missed_steer retention: when the child's result carries pending_steer,
_run_single_child names it on the completion entry (missed_steer field
plus a summary note) so the parent can re-issue the guidance instead of
trusting it landed. This is what makes adding a sender safe: without it
the finish-before-drain race silently loses the text — the exact loss
the finalizer contract exists to prevent.
- subagent.steer gateway RPC beside subagent.interrupt so programmatic
hosts (dashboard, voice layers, ACP bridges) get an in-tree caller;
catalogued in programmatic-integration.md.
- docs: "Steering a Running Subagent" section in delegation.md covering
the queued-vs-delivered semantics.
Tests: registry-level steer coverage (delivery, unknown id, empty text,
dead record, raising agent), the finish-before-drain race retaining
missed_steer, and the RPC contract (4000/4002 validation, queued and
rejected envelopes).
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
/refine [focus] fires the existing background review fork
(AIAgent._spawn_background_review) immediately instead of waiting for
the automatic 10-turn memory / 10-iteration skill nudge counters.
Optional focus instructions are appended to the review prompt so the
fork prioritizes what the user asked for (e.g. '/refine save the
deploy workflow as a skill').
- New optional focus parameter threaded through
_spawn_background_review -> spawn_background_review_thread.
Automatic post-turn reviews pass None and their prompts are
byte-identical to before.
- CLI handler snapshots conversation_history; gateway handler pulls
the idle session's cached AIAgent from _agent_cache (rejected while
the agent is running).
- Review runs in a daemon thread against the snapshot — live
conversation, message alternation, and prompt cache untouched.
- Slack stays under the 50-slash cap via /hermes refine.
Adapted from the /refine concept in Prime Intellect's Prime-Agent
(Continual Harness); Hermes' equivalent durable state is the
memory + skill stores, so the review fork is the natural target.
/heartbeat every <interval> <prompt> gives the current session one
recurring instruction. When the session is idle and the interval has
elapsed, the prompt is injected as a plain user turn — same
conversation, same context, prompt cache and role alternation
untouched.
- CLI: idle-poll watchdog thread (wake-word watchdog pattern) feeding
_pending_input; gateway: single gateway-wide async poller injecting
through the adapter FIFO. Busy sessions coalesce their tick to the
next idle poll.
- Missed ticks coalesce (anchor resets on fire) — a busy hour yields
ONE heartbeat turn, never a backlog. Real user messages always win.
- 60s interval floor; injected prompt carries a don't-invent-work
guard so idle heartbeats don't generate busywork.
- State persists in SessionDB.state_meta (heartbeat:<session_id>),
survives /resume, migrates across compression session rotations
alongside /goal state.
- Session-scoped and in-process by design — durable cross-process
schedules remain the cron subsystem's job (docs draw the boundary).
- Slack stays under the 50-slash cap via /hermes heartbeat; ghost-text
suggester now prefers the shortest prefix match so /he still
suggests /help.
Adapted from the session-heartbeat concept in Prime Intellect's
Prime-Agent (/heartbeat).
/goal gate add <command> attaches shell commands to the active goal.
Gates run at turn boundary BEFORE the LLM judge: a failing gate skips
the judge entirely and feeds its exit code + bounded output tail back
as the continuation prompt, so the agent iterates against concrete
evidence instead of a prose verdict.
- Unchanged-workspace skip: a gate that failed on an identical
workspace (git HEAD + status fingerprint) is not re-run — the
recorded failure replays and the attempt count advances.
- Bounded retries (default 3) + per-gate timeout (default 300s);
exhaustion auto-pauses the goal like the turn budget does.
- Gates persist in SessionDB.state_meta with the goal (survive
/resume and compression rotation); pre-gate goal rows load
unchanged.
- /goal gate [list|add|remove|clear] on CLI + gateway; 'gate' added
to the mid-run control-verb whitelist (gates only run at turn
boundary, so editing the list mid-run is safe).
Adapted from the quality-gate concept in Prime Intellect's Prime-Agent
(--autonomous-gate).
- optional-skills/devops/actual-setup: field-tested setup skill contributed
by shl0ms, updated for the first-class 'actual' provider (the original
targeted a custom-provider config that now collides with the built-in name)
- docs: providers.md section + tables, environment-variables.md, quickstart.md
- tests/skills: frontmatter + first-class-provider conformance checks
* feat(agent): read_preview — the desktop-gated tool that reads the in-app browser
The agent could open the preview pane (open_preview) and read the embedded
terminal (read_terminal), but the browser it had just opened was a black box —
'what does this page say?' had no answer. read_preview mirrors read_terminal
end to end: HERMES_DESKTOP-gated via check_fn (zero schema footprint outside
the GUI), dispatched through the same agent callback pattern, windowed with
start/count so a long page pages instead of flooding context.
* feat(gateway): preview.read blocking bridge
Same lifecycle as terminal.read: the tool blocks on preview.read.request, the
renderer answers preview.read.respond (allow_expired — a slow page extraction
losing the 45s race must not surface a raw 4009), and a timeout emits
preview.read.expire so late answers resolve quietly.
* feat(desktop): the renderer serializes the active preview tab for the agent
preview-reader.ts is the preview analog of the terminal's buffer registry: the
URL pane registers a page reader (webview executeJavaScript → title + visible
innerText) keyed by tab id; readActivePreview resolves the ACTIVE tab, windows
the text (24k cap per read), and answers file/artifact tabs with identity plus
a note pointing at the tool that reads that content directly. The gateway
event handler answers preview.read.request beside terminal.read.request.
* feat(wake): client-capture wake word for remote desktop
Remote headless backends have no PortAudio mic, so "hey hermes" fails even
when openWakeWord is installed. Let the desktop stream 16 kHz int16 PCM via
wake.feed while detection stays server-side.
- wake_word.capture: auto|local|client (+ GUI client_capture prefer)
- WakeWordDetector external_audio queue + feed_audio API
- wake.feed RPC; wake.start/status report capture + frame_length
- Desktop getUserMedia feeder; stop on wake.detected, re-arm after voice
- Docs + unit tests (26 pass in tests/tools/test_wake_word.py)
* fix(wake): address review on client-capture re-arm and feed queue
- wake.status reports effective capture from the armed detector (client vs
local), plus frame_length/sample_rate; GUI status probes prefer client
- Gateway test doubles accept external_audio on start_listening
- Desktop PCM feeder uses a bounded ordered queue instead of dropping frames
while a wake.feed RPC is in flight
- /wake on and status/re-arm paths pass client_capture so remote reattach works
* fix(wake): auto capture keeps the backend mic when one exists
With capture:auto the desktop always preferred client streaming, so a local
desktop with a working backend mic silently switched from PortAudio to
getUserMedia default-device — dropping wake_word.input_device selection
(#74363). A ready backend input now wins under auto; client capture is the
fallback for a preferring surface on a mic-less backend, and capture:client
still forces streaming.
Also removes the dead auto branch (both arms returned local) and lets the
client-feed test skip cleanly when numpy is absent.
* perf(desktop): coalesce wake.feed frames
Sending one 80 ms frame per RPC is ~12.5 gateway calls/s for as long as the
ear is armed. Drain up to 4 queued frames into a single wake.feed payload
(backend feed() already splits long buffers into engine frames) — ~3 RPCs/s
steady-state. Fix the wake.feed size-cap comment (64000 bytes = 2 s, not
0.5 s).
* docs(config): document wake_word.capture in cli-config.yaml.example
---------
Co-authored-by: Andrew <drew@kainotomic.com>
Remote headless backends have no PortAudio mic, so "hey hermes" fails even
when openWakeWord is installed. Let the desktop stream 16 kHz int16 PCM via
wake.feed while detection stays server-side.
- wake_word.capture: auto|local|client (+ GUI client_capture prefer)
- WakeWordDetector external_audio queue + feed_audio API
- wake.feed RPC; wake.start/status report capture + frame_length
- Desktop getUserMedia feeder; stop on wake.detected, re-arm after voice
- Docs + unit tests (26 pass in tests/tools/test_wake_word.py)
The original PR #78453 said 'A job that was mid-run during a restart
resumes according to the attempt policy described in this page.' This
is misleading — the existing docs explicitly state 'Unknown attempts
are audit records and are never automatically rerun.' Corrected to
accurately describe: the mid-run attempt is marked unknown (not retried),
but the job's next scheduled tick fires normally.
- cron: state explicitly that job definitions survive updates, gateway
restarts and reboots (asked directly in #37542)
- mcp: add a Claude Code bridge tip - mcpServers maps to mcp_servers and
hermes import-agent migrates it (the MCP page never says 'mcpServers'
in the client direction; arrivals from Claude Code get no pointer)
- installation: surface loginctl enable-linger in the non-sudo/service
user section where affected users start (currently only on the
gateway page; #43748)
- sessions: document optimize, optimize-storage, repair, recover and
retitle-skills in the CLI reference (shipped in v0.19.1 --help but
absent from the table) and recommend non-destructive optimize before
prune in the db-growth tip
All wording verified against hermes v0.19.1 --help output and the live
pages on 2026-08-04.
Users with Claude Pro/Max, ChatGPT/Codex, SuperGrok or Gemini plans
cannot find what their plan pays for in Hermes in one place (e.g.
#15291, #27228). One comparison table + per-provider notes; cells the
docs do not yet specify are marked 'not currently documented' rather
than guessed.
The four-file map is currently split across the memory, personality and
context-files pages; 'which file is my agent's brain' is one of the most
frequent support questions (e.g. #20245, #29476). One master table, the
frozen-snapshot rule surfaced with a link, and the two canonical mix-ups
answered directly. Content is drawn from the existing three pages.
The goals page never mentions Kanban and the kanban page references /goal
only inside the goal-mode-cards section, so users assume /goal hands work
to the board (see #26116 - /goal is single-session continuation only).
Adds a decision section to goals.md and the inverse note to kanban.md.
Fold ctx.notifyNative into a ctx.os namespace so every way a plugin
reaches outside the app window lives behind one attributed door instead
of accreting one top-level ctx method per capability:
- ctx.os.notify — the native-notification door from the previous commit,
unchanged semantics (plugin kind pref, away-gating, per-plugin throttle).
- ctx.os.openExternal / ctx.os.revealPath / ctx.os.writeClipboard — the
existing window.hermesDesktop bridge capabilities, now sanctioned and
result-shaped: each resolves false (never throws) when the bridge or
member is missing, so a plugin branches on the result instead of
sniffing the preload surface or crashing on an older shell.
No new Electron surface: everything routes through bridge members the
app already ships; the notification path keeps every existing gate.
Desktop plugins can toast in-app (host.notify) but have no sanctioned way to
reach the OS notification pipeline the app's own approval/turn alerts use, so
a plugin surfacing a genuinely notable background event (e.g. a discovery
plugin finding a match) stays invisible once the user steps away from Hermes.
Add a curated per-plugin door instead of exporting the raw dispatcher:
- ctx.notifyNative({ title, body?, silent? }) on PluginContext — attributed
to the plugin id, routed through dispatchNativeNotification so every
existing gate applies (master + per-kind prefs, post-connect baseline,
away-from-app gating, throttle).
- New 'plugin' native-notification kind with its own Settings ▸ Notifications
toggle (default on), so users silence plugins without losing app alerts.
- New optional `tag` discriminator on the notify payload keys the renderer
throttle and main-process cross-window dedupe per plugin, so two plugins
can't collapse each other's session-less notifications.
Consumer: the Index Network desktop plugin wants background opportunity
alerts; anything in ~/.hermes/desktop-plugins gets the same door.
The reset keywords have existed in both CLI and gateway handlers since
June but were undocumented — users couldn't find how to cancel a
personality overlay. Adds a 'Resetting to the default' section to the
personality feature page and mentions the reset in the CLI guide,
slash-command reference (both tables), and messaging command table.
restore_primary_runtime retries the primary every turn once the 60s
transient cooldown clears. For subscription-window limits (Claude
Pro/Max 5h windows, Codex weekly caps) the reset is hours or days away,
so every retry is a guaranteed failure costing two provider switches
and two prompt-cache invalidations per turn.
Add CredentialPool.next_available_at() (earliest reset across exhausted
entries; None when available now or no reset info) and gate the restore
on it: skip while the primary's pool says nobody can serve, restore on
the first turn after the reset elapses. Fail-open: any gate error or
missing reset info falls through to the existing per-turn retry, so
recovery can never be later than today. Cross-provider fallbacks
consult the PRIMARY's pool (not the attached fallback pool), reusing
the loaded pool for the existing rebind to keep auth reads at one per
restore.
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.
The top-level delegate_task description repeated content the model already
receives through parameter descriptions: the concurrency limit (tasks param),
the full nesting clause (role param), context-passing guidance (goal/context
params), and background semantics (background param). Every API call paid for
the duplication (~4,000 chars).
The description now carries only what exists nowhere else in the schema:
use/don't-use routing (execute_code, cronjob), the no-poll rule, the
non-durability warning, the self-report verification contract with concrete
verbs, the language-passing example, the leaf blocked-tool list, and model
inheritance. 3,963 -> 1,704 chars (~570 tokens saved per API call), and the
top-level text is now static (dynamic limits flow only through the two param
descriptions, which are already rebuilt per get_definitions() call).
A/B benchmark across 4 models (gpt-4o, gpt-4o-mini, claude-haiku-4.5,
llama-3.3-70b) showed the naive compaction in PR #72813 regressed weaker
models on exactly the passages it cut (side-effect verification 8/8->0/8 on
gpt-4o-mini; language passing 3/3->0/3 on haiku-4.5). This version keeps
those benchmark-sensitive hooks verbatim.
Tests pin the contracts at keyword level (not prose-literal) plus a size
ceiling, and verify dynamic limits still reach the model via the tasks/role
param descriptions.
Refs #72737, supersedes the delegate_task half of PR #72813.
Live-run findings from a real fact-checking task (ankylosing spondylitis
genetics, 7 authoritative sources) against the new mode:
- Verbatim check rejected a legitimate quote because web_extract returns
markdown: the MedlinePlus sentence is "including _[ERAP1](https://...)_,
_[IL1A](...)_" on the wire but plain prose to a reader. The agent was
forced onto a weaker evidence fragment — the opposite of the point.
Matching now canonicalizes inline links to their label and drops
emphasis/code markers and backslash escapes on both sides, so quoting
the sentence a reader sees works. Paraphrases are still rejected.
- Escaped asterisks (HLA-B\*27) no longer have to be reproduced in the
quote, so extractor artifacts stop leaking into rendered evidence.
- New `render --replace-in <draft>`: rewrites a draft's Sources block in
place, idempotently. Previously the only path was hand-slicing the
file, which also tripped over the emitted heading being `## Sources`
while the prose said "Sources:".
- verify stats: report the provenance total that the percentage is
actually computed from (cited + [unverified], counted once), and print
the line as `info:` instead of `warn:` when nothing is wrong. The old
line printed 17 cited / 2 unverified next to 72%, which does not
reconcile — a sentence can be both.
- SKILL.md documents the emitted heading, --replace-in, and exactly what
counts as a prose sentence for --min-coverage.
7 new tests (47 total) using the real MedlinePlus/Frontiers markup;
6 sabotage runs, all red.
Extends the citation ledger with evidence-backed fact-checking:
- New `quote` subcommand attaches verbatim supporting quotes to a
source; the quote is rejected unless it appears verbatim
(whitespace/case-insensitive) in the fetched page text, so a
paraphrase or misremembered figure cannot masquerade as evidence.
- `verify --evidence` fails a draft whose cited sources carry no
attached quote.
- `render --style evidence` prints each source's quotes beneath its
URL, showing the claim -> source -> exact-text chain.
- `[unverified]` marker declares model-knowledge claims; counts toward
--min-coverage so provenance is declared for every sentence without
forcing fake citations.
- SKILL.md: new Fact-Checking Mode section + pitfalls; version 1.1.0.
- 10 new tests (40 total), all proven live by sabotage runs.
Covers the fact-checking/evidence-transparency half of #28289.
- 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 pooled worker closure captured the caller's live `messages` list and
compress_context explicitly supports plugin/legacy context engines that
mutate that list in place — so after a host timeout, a late engine could
rewrite the live conversation (roles, ordering, persisted content)
concurrently with the resumed turn.
The worker now deep-snapshots the transcript on the worker thread before
any engine code runs; the caller's list object is never handed to pooled
code. Results reach caller-visible state only through the returned value
of an ADMITTED commit (the host discards results on timeout/cancel), and
durable SessionDB mutation was already gated behind the commit fence.
No-op passes map the unchanged snapshot back to the caller's original
list so identity-based no-op detection and flush dedup keep working.
Document the thread-safety contract for context-engine and
memory-provider extension points (they now run on pooled threads) in the
module docstring and the context-engine plugin guide.
Regression: an in-place-mutating engine plus host timeout proves the
caller's live transcript is byte-identical WHILE the worker is still
blocked inside the engine (released only after the assertions).
PR #76354 review, blocking finding 3 / merge gate 3.
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.
- New website/docs/user-guide/messaging/a2a.md: when/where to use A2A
(cross-machine, specialist peers, being callable) vs delegation/kanban
for same-machine multi-agent; enable, outbound tools, inbound surface,
security model, env reference, quick test, troubleshooting. Registered
in sidebars.ts and the messaging index.
- README/DESIGN/plugin.yaml/protocol.py prose updated to name the A2A
v1.0 canonical discovery path /.well-known/agent-card.json (the code
already served both; only the docs lagged).
Builds on the adapter list_channels() hook (cherry-picked from #43545 by
@Guoen0):
- plugins/platforms/simplex: implement list_channels() — enumerates
contacts (/contacts) and groups (/groups) over the live daemon
WebSocket into the channel directory. Returns None when the WS is
down so the directory falls back to session discovery instead of
wiping known targets.
- hermes send --list: merge configured-but-undiscovered platforms into
the listing. Previously a platform configured only via env (e.g. a
fresh SimpleX setup used for outbound sends) was silently omitted,
leaving users guessing at platform names.
- format_directory_for_display(): accept an explicit platforms view and
render empty platforms with a targeting hint instead of hiding them.
- docs: simplex hermes-send section.
Reported by Fedpostoffice on Discord (simplex missing from
hermes send --list; guessed platform names simplex-chat/simplex-relay).
- delivery_id is now generated once per firing and used for both the
X-Hermes-Delivery header and the signed body's delivery_id field —
previously they were two different uuid4s, breaking receiver-side
dedupe as documented.
- 3xx responses are no longer followed: urllib's default redirect
handler converts a redirected POST into a body-less GET, silently
dropping the signed payload. Redirects now log a misconfiguration
warning and count as delivery failure (no retry).
- Docs: receiver-side replay-protection guidance (dedupe on
delivery_id, timestamp freshness window) + redirect semantics.
- Tests: 5xx retry count, redirect-not-followed (sabotage-verified),
header/body delivery_id equality.
The inverse of the inbound webhook platform: hooks.outbound in
config.yaml lists HTTP targets + the plugin-hook events they subscribe
to (on_session_end, subagent_stop, post_tool_call, ...). Each firing
POSTs a JSON payload (same top-level shape as shell hooks' stdin wire)
signed GitHub-style with HMAC-SHA256 (X-Hermes-Signature-256).
Rides the existing hook bus — notify-only callbacks registered on the
plugin manager at the same CLI/gateway/main entry points as shell
hooks. Delivery is fire-and-forget via a bounded queue + single daemon
worker thread, so a dead endpoint can never stall a tool call. Bounded
retries (5xx/conn errors once; 4xx never). secret_env preferred over
inline secret. HERMES_SAFE_MODE skips registration. hermes hooks list
shows outbound targets with signed/UNSIGNED status.
Zero new model tools, zero new subsystems.
Answers and written deliverables that rest on retrieved information now get
inline numbered citations plus a mechanically-rendered Sources list, with a
persistent ledger that makes a hallucinated citation detectable.
- skills/research/grounded-citations/scripts/sources.py: stdlib citation
ledger (add/ingest/list/render/verify) at
$HERMES_HOME/cache/citations/ledger.json, profile-aware, O_EXCL-locked so
parallel subagents sharing a ledger can't collide on ids
- SKILL.md: cite-while-drafting procedure, register-at-retrieval rule,
pitfalls, verification gate
- references/citation-formats.md: per-target placement (markdown, LaTeX/PDF,
docx, pptx, xlsx, wiki, BibTeX handoff to research-paper-writing)
- references/grounding-rationale.md: why numbered ids (ALCE 2305.14627,
WebGPT 2112.09332, Perplexity marker conventions), and how this relates to
the in-process registry in PR #44833
- tests/skills/test_grounded_citations_skill.py: 30 tests
Every ORDER BY id query on the messages table sorted or scanned the
whole session: get_messages_around's window seek, latest_message_row_id
(LIMIT 1), and get_messages' full-load ordering all paid O(session
history) per call — hot mid-turn via session_search and reactions.
messages.id is an original column (INTEGER PRIMARY KEY AUTOINCREMENT),
so the index lives in SCHEMA_SQL next to idx_messages_session — no
legacy-column migration hazard (the kanban lesson from #28776 does not
apply).
Measured (real schema, one 20k-message session, median of 30):
get_messages_around 7.08 -> 0.22 ms (32x), latest_message_row_id
3.37 -> 0.011 ms (307x), get_messages full load 111.6 -> 98.6 ms
(1.13x — remaining cost is row deserialization, not the sort).
Window results byte-identical at probe points across the session.
Tests: VM-step pin (get_messages_around bounded work, calibrated
~12 vs ~855 handler calls, threshold 300 — fails without the index)
and window parity with/without the index. No EXPLAIN/plan text
(behavior contracts, AGENTS.md).
Hermes now pins its toolchain to Node 26 everywhere. Every path that
installs, accepts, heals, or upgrades a Node runtime moves from the old
22-default / `^20.19 || >=22.12` floor to a single rule: Node >=26.
Installers:
- scripts/install.sh — NODE_VERSION=26; node_satisfies_build() collapses
the two-branch Vite floor to `major >= 26`; user-facing messages updated.
- scripts/install.ps1 — $NodeVersion=26; Test-NodeVersionOk likewise;
winget fallback switches OpenJS.NodeJS.LTS -> OpenJS.NodeJS (26 is
Current, not LTS — the LTS manifest would reinstall a too-old Node).
- Dockerfile — node_source stage node:22-bookworm-slim -> node:26 (digest
pinned, amd64 sha256:9e6f...bf73).
- nix/ was already on nodejs_26 (lib.nix, npm-12-0-2.nix); the checks.nix
wrapper check ratchets from `>= 20` to `>= 26`.
Heal/upgrade paths:
- scripts/lib/node-bootstrap.sh — HERMES_NODE_TARGET_MAJOR default 22->26
and HERMES_NODE_MIN_VERSION default 20->26, so heal_managed_node,
_nb_install_bundled_node, and the fnm/proto/nvm/brew rungs all target 26
and stop accepting an on-PATH Node below it. Both remain env-overridable.
- hermes_constants.py — _HERMES_NODE_TARGET_MAJOR fallback 22->26, which
drives the Windows heal path's latest-v26.x download.
Version gates:
- package.json engines.node >=20 -> >=26; apps/desktop engines
`^20.19.0 || >=22.12.0` -> `>=26.0.0`.
- CI setup-node: all five workflows 22 -> 26.
- Docs describing Hermes's own toolchain updated (windows-native, docker,
acp, nix-setup, contributing). Skill docs describing third-party tools'
own requirements are untouched.
Termux still installs via `pkg install nodejs` best-effort (nodejs.org
ships no Android tarballs); that path was never version-gated.
Verified: bash -n on both shell scripts, PowerShell AST parse of
install.ps1, latest-v26.x index resolves (node-v26.5.1), and the install
test suite — 18 tests across the 5 install/runtime test files — passes.
Follow-ups from sweeper review of #43763:
- tests/docker/test_tini_compat_shim.py asserts the dispatcher
ENTRYPOINT (with /init delegation check) instead of a bare /init
- tests/docker/test_smoke.py gains a docker run --init regression
for the non-PID-1 fallback (#38349)
- website/docs/user-guide/docker.md and the s6 supervision skill
document the dispatcher and its wrapped-runtime fallback
Google API and authentication packages permit vulnerable httplib2 and pyasn1
transitives, while the Workspace and Google Chat runtime installers previously
treated any importable version as sufficient. Existing environments could
therefore remain vulnerable after the project dependency pins were repaired.
Carry the fixed versions through the Google and Vertex extras, lazy feature
requirements, lockfile, and both runtime installers. Route the documented
Google Chat installation path through its maintained secure requirements
instead of an unconstrained direct pip command.
Detect stale distributions, install only unsatisfied requirements, and verify
the result before continuing. Behavioral tests cover those repair invariants
without freezing manifests, lockfiles, or complete package sets.
Related #72108
Extracted from #72840
Co-authored-by: Teknium <127238744+teknium1@users.noreply.github.com>
Follows PR #75592: the notifier is no longer gated on
kanban.dispatch_in_gateway. Every gateway delivers events for
subscriptions owned by the profiles whose adapters it hosts; legacy
unstamped subscriptions go only through the confirmed dispatcher
lock owner. Adds a 'Multi-profile setups' subsection to the kanban
Gateway notifications docs (en + zh-Hans).
Finishes the input stash started in the preceding commit from PR #4771.
That PR shipped only the panel renderer: its `@kb.add('c-s')` handler and
stash-state initialization were lost in a rebase, so the panel predicate
read undefined `_stash_panel_open` / `_stash_list` and the feature was
unreachable. This adds the missing half and the tests the PR never had.
Resolves the review feedback on #4771:
- Rebuilt the stash on current main's keybinding setup. The `c-s` key was
unbound repo-wide, so there is no conflict.
- Extracted the state machine into `hermes_cli/prompt_stash.py` as pure
functions (no prompt_toolkit import) so it is directly unit testable —
the PR was cli.py-only with zero tests.
- Dropped the PR's unrelated changes: delegation `supervisor_model` /
`execution_model` config aliases, and stale reverts of the banner
builder, worktree pruning, logging setup, and MCP toolset validation
that its 14k-commit-old base dragged along.
- Fixed the 📌 double-width measurement for real. Three commits in the PR
("subtract 1 from len()", "use bare len()", "subtract 1 again") were
chasing this by tweaking `len()`; all horizontal math now goes through
`_status_bar_display_width` (prompt_toolkit `get_cwidth`), which also
keeps CJK previews inside the border. Narrow terminals fall back to
compact header/footer labels instead of overflowing — caught by a
parametrized width test, not by eyeballing.
Gesture (the contributor's design, kept):
- Composer has content → push onto the stash, clear the input.
- Composer empty, one stashed → pop it straight back.
- Composer empty, 2+ stashed → open the browse panel (↑↓ / Enter / D / Esc).
- Panel open → Ctrl+S closes it.
Pushing onto a stack rather than a single slot is what makes repeated
Ctrl+S safe: a second stash never silently overwrites the first, and with
2+ parked the panel asks rather than guessing which to restore. A `📌 N`
status-bar badge and a composer placeholder advertise the parked draft so
it cannot be silently forgotten.
Deliberate departures from the PR:
- No auto-restore after the agent responds, and no `display.stash_auto_restore`
config key. The PR itself had already defaulted this to false as
"avoids surprising the user"; a keystroke the user pressed should not
cause text to reappear on its own, so the dead default is dropped
rather than carried as config surface.
- Nothing is persisted to disk. Drafts routinely contain pasted
credentials and NDA material, so the stash is session-scoped and
in-memory only. Any future persistence must route through
`get_hermes_home()`.
- Suppressed while a modal prompt owns the composer (sudo / secret /
approval / clarify / slash-confirm / model picker) so Ctrl+S can never
stash a password.
- Restoring images extends `_attached_images` instead of replacing it, so
an attachment added since the stash was taken is not silently dropped.
- `buf.reset()` on stash (not `text = ""`) clears completion state,
selection, and undo stack with the text.
Tests: 95 new tests across two files — 66 on the state machine (empty
buffer is a no-op, exact round-trips including newlines/tabs/CJK/fences,
no-clobber ordering, cap eviction, indicator states, panel cursor
clamping and deletion, the full resolve_ctrl_s decision table) and 29 on
the cli.py wiring (per-instance stash, keybinding registration guard,
layout slot, panel bounded at 8 widths, status-bar indicator lifecycle).
The keybinding-registration test asserts the `c-s` handler exists in
source specifically so the rebase loss that broke #4771 cannot recur.
Verified: 153 passed, 0 failed across the two new files plus
tests/cli/test_cli_init.py and tests/cli/test_cli_extension_hooks.py.
ruff check clean; check-windows-footguns clean.
Docs: Ctrl+S added to the CLI keybindings table.
Co-authored-by: CK iRonin.IT <cyprian@ironin.pl>
The test runner moved to per-file subprocess isolation via
scripts/run_tests_parallel.py (hermetic `env -i`, worker count auto-scaled
from CPU count, FLAKY-retry policy) — no pytest-xdist, no SIGALRM per-test
timeout fixture. Docs still described the old runner in many places:
- AGENTS.md: "-n auto xdist workers, in-tree subprocess-isolation plugin"
clause replaced with the current per-file-subprocess description; the
`::test_x` single-test example now shows file + -k (runner is
file-granular).
- CONTRIBUTING.md: "hermetic env, 4 xdist workers" comment corrected;
`tests/conftest.py::_enforce_test_timeout` reference redirected to the
win32 timeout-method shim in `tests/conftest.py::pytest_configure`.
- skills/autonomous-ai-agents/hermes-agent/references/contributor-guide.md
and windows-quirks.md: same corrections (the bundled skill mirrors the
contributor docs); Windows workaround no longer installs pytest-xdist
or passes -n 0.
- website/docs + zh-Hans i18n mirrors: same fixes in adding-providers.md
and the bundled-skill doc pages.
- skills/software-development/python-debugpy/SKILL.md (+ zh-Hans mirror):
"-p no:xdist"/"-n 0" pdb advice rewritten for the captured per-file
subprocess runner.
- skills/creative/comfyui/tests/README.md: parent-repo "-n auto by
default" rationale updated to past tense.
Combined salvage of PR #38295 (konsisumer), PR #51354 (TutkuEroglu,
redirected to the current conftest truth and the relocated
references/contributor-guide.md), and PR #54956 (waroffchange).
Co-authored-by: TutkuEroglu <rrandqua@gmail.com>
Co-authored-by: waroffchange <116298975+waroffchange@users.noreply.github.com>
Buzz Desktop v0.5.1 now renders Hermes' ACP model menu in agent runtime
settings. Add a short note under the Buzz Desktop host section explaining
where the list comes from (the shared authenticated-provider inventory),
the provider:model / custom:<name>:<model> ID shapes, and that a pick is
session-scoped rather than a Hermes-wide default change.
Replaces the half-duplex per-playback barge monitors with ONE listener
that runs for the entire agent turn in continuous voice mode: armed at
utterance-submit, disarmed when the turn is fully done (response + TTS
finished). Fixes Teknium's live report that voice interruption never
works: (a) not while the LLM is generating, (b) not while TTS plays.
Root causes:
- HALF-DUPLEX GAP: the barge monitor only spawned when TTS playback
STARTED (cli.py streaming/whole-file paths, gateway _tts_stream_begin).
During LLM generation there was NO microphone listener at all.
- PLAYBACK DEAFNESS: the monitor calibrated its VAD noise floor WHILE the
speaker was blasting TTS (speaker bleed baked into the floor), then
multiplied it by 8x with a 1s strictly-consecutive block requirement —
normal speech could rarely reach the trigger, and the 2s grace
swallowed early interjections.
New model — tools/voice_mode.full_duplex_listen():
- Pre-playback calibration: quiet-room noise floor established at turn
start and HELD through playback (never recalibrated against bleed).
- Phase-aware trigger: generation = floor x voice.barge_in_threshold_multiplier
(new config, default 3.0, justified by synthetic-frame tests);
playback = additionally clamped to a 1500-RMS minimum so bleed alone
can't trip; 4000-RMS ceiling keeps speech always reachable.
- Windowed-majority detection (>=80% of a 300ms window) instead of the
strictly-consecutive counter that reset on intra-word energy dips.
- Grace on playback ONSET only (voice.barge_in_grace_seconds, default
down 2.0 -> 0.5) — suppresses the onset transient, not the mic.
- Debug diagnostics at every decision point (calibrated floor, per-window
RMS above 50% of trigger, trip/no-trip, grace suppressions) — always
logger.debug, mirrored to stderr under HERMES_VOICE_DEBUG=1.
Phase behavior (CLI cli.py + tui_gateway/server.py, same model):
- generation: speech interrupts the in-flight turn via the SAME seam the
typed/Ctrl+C interrupt uses (agent.interrupt()), cuts any pending TTS
pipeline so the stale reply never plays, and submits the captured
interjection (pre-roll capture, first syllable kept) as the next turn.
- playback: cuts TTS (streaming pipeline stop + fallback speak stop
events + file player) and submits the capture.
- stop phrase honored in BOTH phases: mid-generation 'stop' interrupts
the turn AND ends the voice chat (stop everything).
- one listener instance spans generation -> playback (no re-arm race);
double-arm refused (CLI _voice_fd_active / gateway _fd_listener_active).
Gateway specifics: _arm_full_duplex_listener() at _run_prompt_submit turn
start and inside _tts_stream_begin; _speak_text_with_barge registers its
(stop, done) pair in _fd_speak_pipelines so fallback speaks are cut and
tracked; _tts_stream_barge_in_monitor kept as a shim that arms the new
listener. Desktop renderer owns its own mic path (voice-barge-in.ts) and
is unaffected; if desktop backend-mic mode is used it inherits via the
gateway.
Tests: full_duplex_listen synthetic-RMS suite (speech-over-bleed trips,
bleed alone doesn't, quiet floor held through playback, grace window,
multiplier math 3x vs 8x, windowed-majority dips), CLI listener phase
tests (generation interrupt seam, playback cut, lifecycle spans phases,
double-arm, config forwarding, stop-phrase-mid-generation), gateway
generation-phase interrupt + stop-phrase tests. Generation-interrupt
test sabotage-verified.
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)
Follow-ups on top of @DonutsDelivery's salvaged reaper commit (#64141):
- create_from_config() now parses lsp.idle_timeout (invalid values fall
back to DEFAULT_IDLE_TIMEOUT) — previously the constructor knob was
unreachable from config.yaml (config exposure adapted from #36892 by
@0xbWy and #68091 by @9miya20)
- canonical default declared in hermes_cli DEFAULT_CONFIG so config
discovery surfaces the knob (per sweeper review note on #64980)
- reaper loop survives transient sweep errors instead of dying and
silently re-opening the leak (gap flagged in #68091 review)
- eventlog.log_reaped(): one INFO line per sweep + clears the
log_active announce cache so respawns re-announce at INFO
- docs: replace the stale 'no idle-timeout reaper' paragraph with the
new lifecycle description + config reference
- tests: reuse-refresh protection (the regression teknium's sweeper
requested on #64141), reaper-survives-error, config propagation,
invalid-value fallback, DEFAULT_CONFIG/manager-constant sync
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.
Saying OR typing a configured stop phrase (voice.stop_phrases, default
"stop") now ends the voice chat everywhere, not just classic CLI PTT:
- hermes_cli/voice.py: new explicit on_stop_phrase callback through
start_continuous/stop_continuous. The force-transcribe path previously
DISCARDED the stop phrase silently — with auto_restart=False the client
re-arms the next capture, so the conversation never ended. Both halt
paths now fire on_stop_phrase (fallback: on_silent_limit for legacy
callers) as user intent, distinct from the no-speech timeout.
- tui_gateway/server.py: voice.record wires on_stop_phrase and emits
voice.transcript {stop_phrase: true} after flipping HERMES_VOICE(_TTS)
off and stopping streaming TTS — same teardown as /voice off. The TTS
barge-in monitor stop-checks its transcript too. prompt.submit consumes
a TYPED bare stop phrase at the server-side choke point when voice mode
is on (returns {voice_stopped: true}, no turn starts).
- ui-tui: voice.transcript {stop_phrase} ends voice mode with a clear
'voice chat ended' notice (distinct from the no-speech-limit message);
submitPrompt releases the busy latch on a consumed voice_stopped reply.
- cli.py: _typed_voice_stop in process_loop — typing a bare stop phrase
while voice mode/continuous is active ends voice mode instead of
sending 'stop' to the agent; typed 'stop' outside voice mode is
unchanged. Voice transcripts skip the check (already stop-checked).
- desktop: interceptsTypedVoiceStop — the composer's onSubmit ends the
live voice conversation (same path as clicking end on the pill) when a
bare stop command is typed with no attachments; renderer-owned loop, so
handled client-side like the existing spoken isVoiceStopCommand.
- tools/voice_mode.py: transcribe_recording never lets the Whisper
hallucination filter swallow a configured stop phrase (e.g. 'bye'
configured as a stop phrase is both a hallucination-blocklist entry and
a stop phrase — stop-phrase check now wins).
Tests: continuous-loop signal (sabotage-verified), force-transcribe stop
signal + legacy fallback, hallucination-filter ordering, typed-stop CLI
unit tests (voice on/off/longer text), prompt.submit typed-stop gateway
tests, TUI vitest for stop_phrase event handling, desktop vitest for the
typed-stop interceptor.
The Skills Hub 'Marketplace' tab showed a single useless entry: Anthropic
changed .claude-plugin/marketplace.json to bundle-shaped plugins whose
source is './', so all plugins collapsed to one identifier pointing at the
repo root, and the second marketplace repo (aiskillstore/marketplace) is
gone (404). Everything in anthropics/skills is already surfaced by the
GitHub tap as the Anthropic tab, making this source fully redundant.
Removes ClaudeMarketplaceSource and all wiring: source router, index
builder (crawl + floors + sort order + rate-limit messaging), extract
labels/install/URL mapping, hub UI tab, web server labels, CLI limits,
docs (en + zh), the legacy index-cache snapshot, and test fixtures.
Stale skills-index entries with source 'claude-marketplace' still install
fine: HermesIndexSource fetches via resolved GitHub paths generically.
One page consolidating all three Hermes×Buzz integration paths —
Desktop managed runtime, buzz-acp relay bridge, and the native gateway
platform — with a comparison table, per-path pointers into the detailed
docs, identity guidance, and contributor credits. Registered in
sidebars.ts under Integrations; Buzz added to the messaging platform
list and a new Collaboration Workspaces section on the integrations
index. en + zh-Hans.
LINE plugin.yaml plus line/wecom-callback/msgraph-webhook/
whatsapp-cloud/teams user-guide pages still documented the old
IPv4-only 0.0.0.0 defaults; update to the dual-stack unset default.
Telegram webhook env docs live in the adapter docstring (updated with
the code change); its plugin.yaml has no webhook host entry.
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.
The multiplexed listener now rejects the default API_SERVER_KEY on
/p/<profile>/ prefixes (fail-closed per-profile keys). Add the
multi-profile routing section with an explicit breaking-change callout
for the next release notes.
Consolidates the native-transport half of PR #73636 by @ScaleLeanChris
onto the merged adapter: persistent NIP-42-authenticated Nostr WebSocket
subscription as the default inbound path (transport=auto|websocket|poll),
kind-44100 membership events for live DM discovery, since-timestamp
resume on reconnect with bounded exponential backoff, and automatic
fallback to CLI polling when the WS can't be established. Events route
through the same _handle_event() pipeline as the poll loop, so de-dupe,
mention gating, p-tag DM latching, and allow-lists behave identically on
both transports. Outbound stays on the CLI (one-shot sends never race a
WS auth handshake — his design).
E2E verified against a real in-process websockets relay: NIP-42
challenge -> signed kind-22242 AUTH (event id re-derived server-side) ->
REQ subscription -> EVENT dispatch -> clean disconnect.
Co-authored-by: ScaleLeanChris <chris@scalelean.com>
The sidecar auth token is generated at spawn (secrets.token_hex) and
existed only in the gateway process memory + sidecar child env, so
_standalone_send from cron subprocesses, hermes send, or the dashboard
structurally could not authenticate (#69960).
The adapter now writes <hermes-home>/runtime/photon-sidecar.json
({port, token, pid}, 0600, atomic tempfile+os.replace) once the sidecar
passes its /healthz readiness check, and deletes it in _stop_sidecar,
on every startup-failure path, and at disconnect so a stale record
never outlives a dead sidecar. _standalone_send falls back to the
record when PHOTON_SIDECAR_TOKEN is unset, validating the recorded pid
is alive first; a stale record yields a clear 'gateway appears to be
down' error. Docs note the gateway-must-be-running requirement and the
Photon-side shared-line initiation policy (#51897).
Command providers legitimately reference their own API keys in shell
templates (curl one-liners). The #70342 scrub removes ALL provider keys,
which would break such setups. Add a per-provider env_passthrough list
(TTS + STT) that copies named variables back from the parent env, plus
docs and tests. Scrub stays the default; passthrough is explicit opt-in.
HERMES_LOCAL_STT_COMMAND rendered quoted placeholders into a
user-configured template and passed the result to shell=True. Shell
metacharacters in the template therefore remained executable syntax even
though the placeholder values themselves were quoted.
Tokenize the rendered template and invoke it as an argv list while
preserving the existing timeout, closed stdin, and Windows creation flags.
Lock the invocation contract with metacharacter regression coverage and
document explicit shell wrapping for trusted templates that need it.
Salvages #32694
Co-authored-by: Ernest Hysa <takis312@hotmail.com>
Command-type TTS providers validated output_format against a hardcoded
{mp3,wav,ogg,flac} set; any other value was silently coerced back to mp3,
which then mismatched the output path the post-run check expects. This
blocked common ffmpeg-producible containers/codecs — notably m4a (AAC),
the portable choice for WeChat/iOS/mobile voice files — with no
config-only path (only a local source patch, lost on every update).
Widen COMMAND_TTS_OUTPUT_FORMATS to add m4a, aac, amr, opus. This only
permits a command provider to declare these; the user's command still
produces the file (e.g. via ffmpeg). No built-in provider behavior
changes and no new required config.
Update the two tests that pinned the old set, and add a positive case
covering the new formats. Document the supported output_format values.
Expands the 'Recommended display settings' section into a comprehensive
'Recommended default settings' block covering display, access control,
polling, and mention behavior. Each setting includes an inline annotation
and rationale bullet. Matches Telegram/email default behavior (no
intermediate tool output, mention-gated channels, private-by-default
access).
Refs #68871
Plugin-path adapter (zero core changes) connecting Hermes to a Buzz
community relay via the buzz CLI binary (JSON in/out, arg-list exec,
key passed via env only). Inbound uses a poll loop with per-channel
high-water marks seeded from newest (no history replay), event-id
de-dupe, self-echo suppression by pubkey, and mention gating in
channels (DMs always dispatch). Registers env_enablement, cron
home-channel delivery, and an out-of-process standalone sender,
mirroring the IRC plugin.
Verified against a live relay: connect -> send -> poll -> MessageEvent
round-trip, self-echo suppressed, clean disconnect.
Known limitation: polled inbound (default 4s); a websocket transport
(buzz-ws-client) is a future optimization.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Follow-up on the salvaged #69915 commits:
- verified env vars against block/buzz crates/buzz-acp/src/config.rs
(BUZZ_RELAY_URL, BUZZ_PRIVATE_KEY, BUZZ_API_TOKEN, BUZZ_ACP_AGENT_COMMAND/
ARGS, BUZZ_ACP_RELAY_OBSERVER, BUZZ_ACP_AGENT_OWNER) and kind 24200
observer frames against docs/nips/NIP-AO.md
- dropped unverifiable claims: docs/hermes-agent-acp.md link (404 upstream),
relay-directory profile publication / External-agents card walkthrough,
'bot role' membership phrasing
- replaced key provisioning with the actual buzz-admin generate-key /
add-member flow from the buzz-acp README
- retitled the section 'Buzz channels (relay bridge)' and cross-linked it
with the Buzz Desktop managed-runtime section both ways; permission
paragraph now points at the owner-only warning instead of duplicating it
- zh-Hans translation of the new section + retitle to ACP 宿主集成
* 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.
Teknium
mzkarami
kshitij
Jack Maloney
Brooklyn Nicholson
Ken Weiner
kinsolee
GodsBoy
HexLab98
kshitij
texasich
xxxigm
SmokeDev
Andrew
witcheer
seref
konsisumer
EndeavorYen
Carbon
WojtekMR3
Kyzcreig
fangliquanflq
fangliquanflq
Hermes agent
spfcraze
ethernet
Zeraphim
Eugeniusz Gilewski
Gille
Francesco Bonacci
virtuadex
Hermes Agent
Sherman
CleanDev-Fix
Rob Zehner
nytemodeonly
Siddharth Balyan
Atakan
3ASiC