The a2a invariant test skipped any synthesized token the canonical
redactor itself let through, so a boundary/shape regression would have
passed silently. Every class scrubs today, so the escape hatch goes and
the soft ">= 50" count becomes the exact registry size.
The gateway body test still asserted a "[REDACTED]" fallback marker that
no longer exists; redact_for_egress masks via _mask_token ("***"), so
assert that alone. honcho oauth.py's `re` import became unused when its
private pattern list moved to the registry.
plugins/platforms/a2a/security.py::redact_outbound shipped text to a REMOTE peer
through 8 private regexes (sk-, sk-ant-, ghp_ only, xox[bap] only, AKIA, JWT,
Bearer, email) and never called redact_sensitive_text, so every prefix added to
agent/redact.py (hf_, glpat-, xapp-, npm_, Telegram bot tokens, private keys,
DB URLs, env assignments, auth headers, plugin-registered patterns) was absent
on the A2A path. gateway/run.py::_GATEWAY_SECRET_PATTERNS and
agent/monitoring/redaction.py::_TOKEN_RE/_BEARER_RE were two more parallel
"fallback" lists to maintain.
Now agent/redact.py::redact_for_egress is the one egress scrub:
redact_sensitive_text(force=True) + a bearer sweep for prefix-less opaque
tokens, fail-closed ("[redaction-unavailable]"). Gateway user-facing text,
monitoring export and A2A outbound call it; A2A keeps only its e-mail pass.
Behavior changes: a2a egress now masks the full canonical set; the gateway
chat path returns the fail-closed sentinel instead of a raw string when the
redactor raises; honcho plugin registers hch-at-/hch-rt- with
register_redaction_patterns (masked on every surface; mask shape is the
shared head/tail form instead of "hch-at-[redacted]"); proxy_cli token
display uses mask_secret (4 visible prefix chars instead of 12).
Invariant test: redact_outbound masks a synthesized token for every
registered prefix pattern (fails when reverted to the private list).
fal launched H3 Max Turbo on Sep 3 (minimax/h3-max-turbo/{text,image}-to-video):
a throughput-tuned post-train of H3 Max with a 1080P tier Max lacks, at
$0.025/s 480p / $0.04/s 768p / $0.08/s 1080p list ($0.00625-0.02/s promo until
Sep 14). Schema matches Max's shape — required prompt_expansion_mode static
key, int duration 5-15, seed on both endpoints, i2v drops aspect_ratio — plus
the new 1080P resolution enum, so it reuses the existing family capability
flags with a Turbo-specific resolution alias map.
Schema verified against the FAL queue OpenAPI for both endpoints. Live E2E
blocked by the FAL account balance lock (403 "Exhausted balance"); portal
allowlist/pricing needed for managed users on the 2 new endpoints.
The deepseek plugin is bundled and always loads, so the try/except around
the delegation import was defense-in-depth for a path that cannot fail.
Tests reduced to the two contracts that matter: /reasoning none reaches
the wire as thinking.disabled, and DeepSeek ids produce exactly the native
DeepSeek profile's output while non-DeepSeek families stay a no-op.
The bundled-plugin loader pops half-registered modules when a plugin
fails to load, so the lazy 'from plugins.model_providers.deepseek import
deepseek' could raise ImportError on every DeepSeek-routed CommandCode
turn — turning the soft 'thinking uncontrollable' bug into a hard
crash. Catch ImportError, log, and return the pre-fix no-op (review
feedback on #95241).
CommandCode fronts DeepSeek with vendor-prefixed ids
(deepseek/deepseek-v4-flash). DeepSeek V4+ defaults to thinking mode
when the thinking field is omitted, so /reasoning none changed the
Hermes session state but not the actual request -- the turn sat in
reflecting.../brainstorming... for minutes (#95232). Strip the vendor
prefix for DeepSeek-family ids and delegate to the native DeepSeek
profile's build_api_kwargs_extras (extra_body.thinking +
reasoning_effort mapping); other CommandCode model families keep the
base no-op behavior. The prior no-op tests codified the bug and are
rewritten to pin the new contract.
The salvaged #103267 plugin hardcoded a single model (minimax/hailuo-3-max) and
rejected any other id. OpenRouter's public GET /api/v1/videos/models already
publishes every generative model with its supported durations, resolutions,
aspect ratios, frame-image support, audio and seed flags, and pricing SKUs, so
the provider now reads that catalog (5-min TTL, offline snapshot fallback):
- list_models(): all 25+ generative models (edit/upscale/avatar rows that take
no duration are outside the unified video_generate surface and are dropped)
with a per-second price label where the SKU is per-second
- capabilities(): the CONFIGURED model's surface, so the dynamic schema only
advertises audio/seed/resolutions the selected model honours
- _build_payload(): clamps duration/resolution/aspect ratio to the model's
live limits (nearest by value/height/ratio) and drops generate_audio/seed
for models that lack them (the API 400s otherwise); reference images ride
in input_references; local file inputs are refused (OpenRouter fetches
URLs itself), data:image/ URLs from the sandbox chokepoint pass through
- bearer key only ever goes to the configured origin (poll + /content),
never to a provider-supplied unsigned_urls host (kept from #103267)
Also drops the source-grep `_IGNORES_SEED` escape hatch #103267 added to the
declaration⇄implementation sweep; the provider now implements seed for real.
Docs list OpenRouter and DeepInfra as bundled video backends.
Requested by Don Piedro Savastano (Discord): OpenRouter credit for video_generate.
The version-less canonical Flash id is not matched by _is_deepseek_thinking_model, so agent.reasoning_effort and every auxiliary reasoning_effort were silently dropped on the OpenCode Go relay while the direct provider was fixed (8435a3ae00/aeecb110f8). Match it through a version-less id set, mirroring plugins/model-providers/deepseek. Verified against the Go relay: named levels are graded (low 906 / high 1371 / max >=2500 reasoning tokens on a multi-step prompt) and integer efforts are rejected, so the named level is the knob to send.
_truncate_for_sync documents "the last sentence boundary within max_len", but it
looped over separator KINDS and returned on the first kind that qualified. An early
"。" therefore outranked a "." 240 characters later, and in pure ASCII "." outranked
a later "!" or "?" purely because it comes first in the tuple.
With the 450-char default, "a"*200 + "。" + "b"*240 + "." + "c"*100 kept 201 of the
442 characters available: 241 characters the embedder would have accepted were
discarded, so any fact in the second half of the turn never reached extraction. The
add() call succeeds, so unlike #106235 nothing is logged — the turn is simply
remembered from its first sentence. Raising sync_max_chars widens the gap rather
than closing it.
Take the max over every separator instead, from a named tuple so the set is not
buried in the loop. ".\n" is dropped: its index can never exceed the bare "." it
starts with, so under a max it is unreachable. The first-third guard and the hard-cut
fallback for unsegmented input are unchanged.
Fixes#108868
Ramp Router efforts cache + warm/disk flags, xAI and OpenRouter image catalogs,
Hindsight append-capability verdict, memory-provider skill registry, OpenViking
atexit provider, Honcho loopback flow status, Langfuse client (os.environ-only
credentials) and disk-cleanup's protected cron paths held one profile's
credential- or home-derived value process-wide; YuanbaoAdapter._active_instance
was last-connected-wins across profiles.
Keyed by home key / credential fingerprint under an override, credentials read
through the secret scope, warm threads run under copy_context(); unscoped module
slots stay for the single-profile path and the existing monkeypatch tests.
Keep: scopeless worker resolves the on-disk key (durability core), a keyed
build still rewrites (rotation), and the API-prefixed vault name resolves.
Drop the mock-patched gate predicate test, the inlined copy of the same
predicate, and the empty-build/empty-disk arm — they re-assert the helper's
shape rather than a behaviour contract.
Under gateway.multiplex_profiles, os.environ holds the DEFAULT profile's .env; a
secondary profile's values exist only in the per-turn secret scope. Every reader
below still read os.environ/os.getenv at call time, so a secondary profile's turn
silently used the default profile's value.
Credentials (F6): FIRECRAWL_API_KEY (read_file hosted OCR), OPENVIKING_API_KEY,
mem0-OSS OPENAI_API_KEY, MODAL_TOKEN_ID/SECRET and BROWSER_USE_API_KEY presence
gates, and the xAI video plugin's os.getenv("XAI_API_KEY") fallback AFTER the
scoped resolver had already missed — the exact fallback-after-miss shape
gateway/AGENTS.md forbids. Deleted, not re-scoped: the resolver is the scope.
Identity / tenant (F7): MEM0_USER_ID/AGENT_ID/HOST/MODE, SUPERMEMORY_CONTAINER_TAG,
RETAINDB_PROJECT, OPENVIKING_ACCOUNT/USER/AGENT (and the whole layered() env
read), HINDSIGHT_BANK_ID/MODE/retain shaping, HERMES_HONCHO_HOST. A raw read
put a secondary profile's memories into the default profile's account/bank/
project/tenant and recalled them back into the default's turns. Each now uses
get_secret with the provider's own per-profile default on a miss.
Endpoints (F8): OPENAI_BASE_URL (aux custom runtime + direct-alias expansion),
XAI_BASE_URL/HERMES_XAI_BASE_URL (aux OAuth), NOUS_INFERENCE_BASE_URL (#65941,
both the aux builder and hermes_cli.auth_nous._nous_inference_env_override),
GATEWAY_PROXY_URL (same UnscopedSecretError-only fallback shape as
GATEWAY_PROXY_KEY three lines below), FIRECRAWL_API_URL, BROWSERBASE_BASE_URL,
SUPERMEMORY/RETAINDB/HONCHO/HINDSIGHT URLs. The keys beside them were already
scoped, so a secondary's key was sent to the default profile's proxy or host.
Targets / display (F11): WEIXIN_HOME_CHANNEL (message posted into the default's
chat), HERMES_LANGUAGE, and agent/i18n's process-wide lru_cache of
display.language — now keyed by HERMES_HOME.
Outbound webhooks: hooks.outbound[].secret_env resolved from os.environ while
the gateway registers each profile's targets inside that profile's scope, so a
secondary's deliveries were signed with the default's secret or left unsigned.
Agent-cache eviction: _spawn_release_thread started a bare threading.Thread, so
commit_memory_session -> provider on_session_end ran with an EMPTY context. The
thread now runs copy_context() and, for the unscoped housekeeping sweep, enters
the owning profile's _profile_runtime_scope resolved from the session key
(agent:<profile>:...). The pressure batch does the same per key.
session_search (#82903): agent/inline_tool_executors.py::_session_search
forwarded every schema argument except `profile`, so a gateway agent could
never select a named profile's store. Forwarded; the ownership-scoping design
in #87779/#87847 is a separate design call and is not attempted here.
Live repro (/tmp/mux_audit/fix-tool-memory-reads/repro.py): 28 FAIL on
origin/main -> 0 FAIL with this change; 10 new invariant tests red on base.
Fixes#82903Fixes#65941Fixes#99121
Addresses #87779
Co-authored-by: webtecnica <75556242+webtecnica@users.noreply.github.com>
Co-authored-by: Michael Versluis (Berry) <michael@wve.nl>
Conflict resolution on top of main also routes MEM0_MODE / MEM0_HOST / MEM0_AGENT_ID /
MEM0_USER_ID through get_secret: identity and host are .env values like the key, and a
raw environ read hands a secondary profile the default profile's mem0 account.
(cherry picked from commit 7a5863ebafdb193d8ec2f89a7aca309774d6d69a)
The vault item is named HINDSIGHT_API_LLM_API_KEY (matching the daemon env var), while the setup wizard uses HINDSIGHT_LLM_API_KEY. Vault-fed scopes carrying only the API-prefixed name resolved empty, so the scopeless path — and the whole fail-closed machinery — never engaged on real deployments. Accept both names.
Proven end-to-end: hydrated scope now resolves the 73-char key and the rewrite gate passes.
(cherry picked from commit 2ce8c2980188ae8e4a3ed398e14fa8ed89a40644)
test_gate_refuses_keyless_build_against_keyed_disk and test_rewrite_allowed_when_build_carries_key touch no filesystem modes, so the nt skip guard was needlessly excluding them. Addresses review feedback on PR #103957.
(cherry picked from commit 196cf723571c87829be61e17df27869948db00a0)
The embedded daemon-start worker usually runs with no secret scope, so _embedded_llm_api_key resolved empty and both the profile-env compare and the upstream manager merge overwrote the file's good key with emptiness.
- _embedded_llm_api_key: explicit config, then secret scope, then on-disk profile env fallback; UnscopedSecretError treated as empty without touching os.environ (multiplex safety).
- Add _may_rewrite_profile_env gate: refuse rewrite on keyless-build plus keyed-disk, and skip the daemon stop in that case.
- Document key resolution order in plugins/memory/hindsight/README.md.
- Add 5 tests pinning disk fallback, the gate's refuse branch, rotation, keyless providers, and file survival.
(cherry picked from commit 48a4bcd536fa799b666a8d5eaae73c1816ca9e33)
A DM relayed from another Hermes profile ran as a turn in the recipient's
Bot Chat session. sync_turn wrote the bot's words and the recipient's reply
into that session, and before per-author writes they landed under the
human's peer. The human's representation absorbed conversations the human
never had.
The turn context now marks such turns with scope a2a:<bot id>.
sync_turn routes a bot-authored turn into a separate Honcho session keyed
<session>:a2a:<sanitized bot id>, created with the sender bot as its user
peer, and never writes it into the human's session. The key is deterministic
so every turn from the same bot reaches the same session, and it stays
inside Honcho's 100 character session id limit. Recall still reads the
human's session only.
a2aSessions (host block, then root, default true) turns the routing on.
With it off, bot-authored turns are skipped. A bot turn that names no
author id is skipped as well, because nothing can key its session. Human
turns are unchanged.
get_or_create takes a user_peer_id override so the a2a session's roster is
the bot and the assistant, not the runtime human.
DeepSeek retired deepseek-v4-flash on 2026-09-10 (V4.1-Flash release); the API's
model name is now `deepseek-flash` and /v1/models lists only it. Hermes still
folded every non-V-series name onto deepseek-v4-flash, so `/model deepseek-flash`
on the DeepSeek provider was rewritten, then the validator "auto-corrected" it
back against the live listing: "Auto-corrected deepseek-v4-flash -> deepseek-flash"
on every switch.
Retired aliases (deepseek-chat / -reasoner and other fuzzy names) now fold onto
deepseek-flash; the curated catalog, profile fallback list, aux default, goal-judge
hint and pricing snapshot (2026-09-10 off-peak USD) follow the docs. Dated
deepseek-v4-* ids still pass through untouched.
Builds on YipTszkwan's #107126 (earliest fix in the cluster).
DeepSeek's 2026-09 Flash refresh introduced a version-less canonical id:
GET /v1/models now returns `deepseek-flash` (alongside `deepseek-v4-pro`), the
API accepts it directly, and the older `deepseek-v4-flash` is server-side
aliased onto it. Every DeepSeek model-id gate in Hermes keys off the
`deepseek-v<N>` prefix, so the new id silently missed all four:
* DeepSeekProfile.build_api_kwargs_extras classified it as non-thinking and
omitted `extra_body.thinking`. The server then defaults to thinking-on, so
the user's thinking toggle and `reasoning_effort` were quietly ignored.
* `_normalize_for_deepseek` folded it onto `deepseek-v4-flash` (it misses the
V-series regex), so the id a user picked never reached the wire and the
config stored a different model than the picker advertised.
* `DEFAULT_CONTEXT_LENGTHS` fell through to the 128K `deepseek` catch-all
instead of the real 1M window, capping the model at an eighth of its
context before compaction kicked in.
* `_REASONING_STALE_TIMEOUT_FLOORS` had no entry, leaving the stale-stream
detector at its 180s default instead of the 600s reasoning-model floor.
Verified live against api.deepseek.com: `deepseek-flash` answers 200 with
`model: deepseek-flash`, accepts image input (the refresh folds vision into
the Flash model), and the in-between id `deepseek-v4.1-flash` is rejected
with "The supported API model names are deepseek-flash, deepseek-v4-pro".
Adds the id to all four gates plus regression coverage for each site.
Keep the two tests that fail on main without the fix:
- test_fal_common: a keyed managed submit makes exactly one POST (plus the
negative arm: an unkeyed submit still goes through the SDK retry ladder)
- test_image_generation: the 409 BILLING_ERROR body surfaces
`unsupported_pricing_meter` instead of the generic "not yet enabled" text
Dropped from #106484: the duplicate video-plugin billing test (same helper,
same assertion), the `_fal_client = fake` / `import_fal_client` stub churn
and the `tools.lazy_deps` stub — fal-client is installed in CI (`--extra fal`)
so those fixtures were not needed; the `_load_fal_client` no-op fixture on
TestManagedGatewayErrorTranslation for the same reason.
Also drop the redundant `retry_request is None` re-check in
`_ManagedFalSyncClient.submit` — `__init__` already raises when the helper
is missing.
Avoid retrying idempotent managed FAL submissions because the retry can mask the initial billing failure. Surface structured Nous billing diagnostics consistently for image and video paths, with hermetic regression coverage.
(cherry picked from commit 289ce039e9a522dc8016ae4a512214c05d0a8bc0)
A flat 450-char cap fits 512-token embedders (bge-small-zh-v1.5,
all-minilm) but stores only ~5% of the window on 8192-token models
(text-embedding-3-small, jina-embeddings-v3, bge-m3), degrading memory
quality for users those models served fine before truncation existed.
Read `sync_max_chars` from mem0.json once in initialize() (450 default)
and pass it to _truncate_for_sync(). Config over auto-detection: the
Ollama /api/show probe + known-model table proposed in #37427 adds a
network call and a curated list for a number the operator already knows
from their embedder choice; the setup wizard's mem0.json is the plugin's
behavioral-settings surface (no new HERMES_* env var). Documented in the
plugin README and the memory-providers docs page.
Dynamic-cap requirement and measurements (450 OK / 600 -> HTTP 500 on
bge-small-zh-v1.5:f16) by @szicely in #106235.
Refs #37421#106235
Co-authored-by: szicely <140148567+szicely@users.noreply.github.com>
Co-authored-by: liuhao1024 <sunsky.lau@gmail.com>
sync_turn() sent the whole turn to backend.add() untruncated. OSS
embedding models with small context windows (Ollama bge-small-zh-v1.5:
512 tokens) reject the request with HTTP 500, and hosted APIs answer
INPUT_TOKEN_LIMIT_EXCEEDED — in both cases _try() only logs, silently
dropping the turn's memory extraction after long conversations.
Cap each synced message at its last sentence boundary within 450 chars
before ingestion: short turns pass through unchanged, long turns keep a
coherent statement for fact extraction. This replaces the previous
retry-on-error approach, which could not match Ollama's HTTP 500 shape.
Salvaged from #37427 with the test suite trimmed to the one invariant
(oversized turn still reaches a small-context backend, short text
untouched, no breaker failure).
Refs #37421#106235
Co-authored-by: szicely <140148567+szicely@users.noreply.github.com>
Drop the bare "tool.content" pattern (any 400 mentioning tool.content in a
non-list context would be sent through the image-strip path) and the profile
flag snapshot test; the behaviour tests (classifier verdict + proactive
downgrade) already pin the contract.
- Declare `supports_vision_tool_messages=False` and `supports_vision=True` on `opencode_go` provider profile in `plugins/model-providers/opencode-zen/__init__.py`
- Route HTTP 422 errors through `_IMAGE_TOOL_RULES` and add `tool.content.str`, `tool.content`, and `input should be a valid string` patterns to `_MULTIMODAL_TOOL_CONTENT_PATTERNS` in `agent/error_classifier.py`
- Add unit tests for OpenCode Go proactive tool result downgrade, HTTP 422 Console Go classification, and profile capability contract in `tests/run_agent/test_multimodal_tool_content_recovery.py` and `tests/plugins/model_providers/test_opencode_go_profile.py`
Breaks the two import cycles that forced Protocol stand-ins in the F821 sweep, so the two
sites now name the real types.
gateway/platforms/event.py (new leaf): MessageType, ProcessingOutcome, MessageEvent moved
out of base.py verbatim. Their only dependency is gateway.session.SessionSource; base.py
imported helpers.py at module level, so helpers could not name MessageEvent. Now
TextBatchAggregator is typed by the real MessageEvent. 249 importers repointed
(`from gateway.platforms.base import` -> `.event`, preserving each import's layout);
gateway.platforms.__init__ re-exports from .event. The three revert-scheduled PLUGIN-COMPAT
pointers that named these symbols (gateway.slash_commands → MessageType, dingtalk → MessageType,
photon → ProcessingOutcome) and their COMPAT_MANIFEST rows now target gateway.platforms.event.
Docs updated: ADDING_A_PLATFORM.md, adding-platform-adapters.md (en + zh-Hans).
tools/mcp_tool_sampling.py: ElicitationHandler no longer holds a back-reference to its
MCPServerTask (mcp_tool imports sampling, so the task type cannot be named there). It only
ever read owner._pending_call_context, so it takes `call_context: Callable[[], Context | None]`
and MCPServerTask passes `lambda: self._pending_call_context`. The consent call is one
`functools.partial`, run directly or inside the captured Context.
ty on the 11 touched production files vs origin/main: 0 new diagnostics, 14 resolved.
(The one `source: SessionSource = None` diagnostic moves with the class; typing it Optional
exposes ~60 unguarded call sites — separate follow-up.)
Tests: tests/gateway + tests/plugins + tests/tools + touched files, 18,235 passed; the 31
failures reproduce identically on origin/main (macOS /private/tmp, systemd socket,
long-path fixtures, live-service tests).
Move the general-vs-memory hook ownership logic out of the memory collector into
PluginLedgerMixin (_drop_fallback_hooks / _register_fallback_hook) so the collector
and the loader each call one manager method instead of reaching into manager privates.
Hoist hashlib to module scope. Trim the new suite to the three invariant cases
(run-once across load orders, distinct sources not suppressed, re-exported register).
The plugin loader execs sibling modules before the package __init__, so a
module-level `from . import _read_mem0_json` in _setup.py failed against the
empty parent shell and the whole module silently dropped out — every later
`hermes memory setup mem0` died with "cannot import name 'post_setup'".
The test loads mem0 through the real discovery path from a cold sys.modules
and asserts the cached _setup module exposes post_setup (red on base).
Also maps the contributor email for #103078 credit.
Campaign tracker: https://github.com/NousResearch/hermes-agent/issues/104154
Adds plugins/web/perplexity — a keyed-only WebSearchProvider over httpx:
- search: POST https://api.perplexity.ai/search (documented Search API),
search_context_size=low so `snippet` stays description-sized;
results[].snippet -> description, max_results capped at the API's 20.
- extract: POST /sdk/content/snippets — the query-relevant page-excerpt
route behind `pplx content snippets` (the CLI's `content fetch` is
deprecated upstream). web_extract has no query, so the URLs' path words
serve as the relevance query; per-URL `error` entries survive a 200.
- Wired into the same touchpoints as the other keyed vendors: legacy
backend set + credential ladder + availability probe (web_tools),
registry preference walk, OPTIONAL_ENV_VARS, `hermes config`/status/
dump key lists, nous_subscription direct-credential detection, setup
summary, test conftests, docs.
Not a keyless-ring member (Perplexity has no anonymous tier). Related
closed PRs #9192 / #23981 / #45225 predate the plugin ABC.
dingtalk: drop DINGTALK_TYPE_MAPPING/EXT_MAP re-exports. google_chat: card_spec_to_cards_v2 test -> .cards.
matrix: drop module-level MAX_MESSAGE_LENGTH alias (no importers). teams: drop TeamsSummaryWriter re-export
(teams_pipeline/runtime + tests -> summary_writer). wecom: drop WeComStreamExpiredError/STREAM_EXPIRED_ERRCODE/
MAX_INTERMEDIATE_FRAMES re-exports (tests -> .streaming). parallel: drop _get_parallel_client/_get_async_parallel_client
aliases (tests -> _get_sync_client). email: drop stale 'alias' comment (_esecret_int is the only name).
photon: re-remove credential_summary() (shim-only, cb9b7c36f3); its no-leak test now drives print_credential_summary.
ethernet8023: the no-leak test was deleted alongside the function it pinned; the
display-only credential status is live in print_credential_summary. credential_summary
restored byte-identical to BASE so the leak contract is CI-pinned again.
