* feat(moa): expose MoA presets as selectable virtual models
Reconstructed onto current main (PR #46081's base had diverged with no common
ancestor, marking the PR dirty so CI never dispatched). MoA is now a virtual
provider: each named preset is a selectable model under provider 'moa', and the
preset's aggregator is the acting model that answers and calls tools.
Reference models fan out in parallel via a bounded ThreadPoolExecutor (the same
batch pattern delegate_task uses) — all references dispatched at once, collected
when every one finishes, then handed to the aggregator. Output order is
preserved, failures and the MoA-recursion guard stay isolated per reference.
- Removed the old mixture_of_agents model tool and moa toolset.
- Added moa as a virtual provider in the provider/model inventory.
- /moa is shortcut behavior over model selection (default preset / named preset
/ one-shot prompt).
- Dashboard + Desktop manage named presets; presets appear in model pickers.
- Parallel reference fan-out in agent/moa_loop.py with regression test.
* fix(moa): thread moa_config through _run_agent to _run_agent_inner
The reconstructed gateway MoA wiring declared moa_config on _run_agent (the
profile-scoping wrapper) and used it inside _run_agent_inner, but the wrapper
never forwarded it — _run_agent_inner had no such parameter, so the runtime hit
NameError: name 'moa_config' is not defined on the compression-failure session
sync path. Add moa_config to _run_agent_inner's signature and forward it from
both wrapper call sites (multiplex and non-multiplex). Caught by
tests/gateway/test_compression_failure_session_sync.py on CI shard test(4).
* fix(moa): classify moa as a virtual provider in the catalog
The moa virtual provider has no PROVIDER_REGISTRY/ProviderProfile entry, so
provider_catalog() fell through to the default auth_type="api_key" with no
env vars — tripping two catalog invariants:
- test_provider_catalog: api_key providers must expose a credential env var
- test_provider_parity: every hermes-model provider must be desktop-configurable
moa already declares auth_type="virtual" in HERMES_OVERLAYS; consult that
overlay as an auth_type fallback so the catalog reports moa as virtual (no real
credential, no network endpoint). Exempt virtual providers from the desktop
parity union check the same way 'custom' is exempt — derived from the catalog,
not a hardcoded slug, so future virtual providers are covered too.
Collapse the duplicated cold-resume / lazy-watch / create scaffolding into
shared helpers: _deferred_session_record (the live-session dict minus the
agent), _lazy_resume_info (the not-yet-built session.info), _claim_or_reuse_live
(lock + double-checked register-or-reuse), and _schedule_agent_build (the
pre-warm timer). Net -12 lines, three copies of the ~30-key session dict and
the lazy-info block down to one each. No behavior change.
Per review: gating the faster path behind a `defer_build` flag that the
only caller always sends is pointless. Flip it — `session.resume` now
defers the agent build by default for every caller (desktop + Ink TUI);
a caller that needs the agent built synchronously passes `eager_build:
true` (used by the build-race test). The desktop no longer sends a flag.
While verifying the flip, fixed two real parity gaps the deferred path
had vs the old eager (`_init_session`) path:
- `_enable_gateway_prompts()` was never called on a deferred resume, so
approvals/clarify wouldn't route through the gateway prompt callbacks.
- `_start_agent_build` never wired `background_review_callback` /
`memory_notifications`, so a deferred-built session's self-improvement
"💾 …" summary leaked to stdout instead of rendering in-transcript.
Wiring it there also fixes it for `session.create` sessions, which
build through the same path.
ACP is unaffected (it uses its own session_manager, not this RPC); the
Ink TUI already consumes the same lazy `info` shape from session.create
and upgrades on the later `session.info` event.
Switching sessions in the desktop app could freeze the whole UI for
several seconds on heavy, tool-rich chats. Root causes and fixes:
- Cold `session.resume` built the AIAgent (MCP discovery, prompt/skill
build) *before* returning, and the desktop awaits that RPC before it
paints — so the entire switch blocked on the build. Add an opt-in
`defer_build` resume path (the contract `session.create` already uses):
return the full display transcript immediately, register an upgradable
live session, and pre-warm the agent on a short timer. The persisted
runtime identity (model/provider/base_url/api_mode/reasoning/tier) is
restored on the deferred build so it can't drop the provider.
- Nothing bounded how many in-memory agents accumulate; a user who
reconnects often piled up detached sessions for the full 6h TTL. Add a
soft LRU cap (`max_live_sessions`, default 16) that evicts the
least-recently-active DETACHED sessions (no live client) — never a
running, awaiting-input, mid-build, or live-transport one. Reopening
re-resumes from disk.
- On the prefetch-hit cold-resume path, skip rebuilding a throwaway
merged-message array (and its 1000-entry Map) when the prefetch already
painted the exact transcript; the downstream sameMessageList guard
already drops the publish, so it was pure main-thread cost.
The desktop opts into `defer_build` for every non-watch cold resume; the
eager path stays for CLI/TUI and existing callers.
atomic_yaml_write used default yaml.dump which emits indentless
sequences (list items at column 0), while atomic_roundtrip_yaml_update
(ruamel.yaml) emits 2-space-indented sequences. Cross-path writes to
the same config.yaml toggled indentation on every save, eventually
producing a mixed-indent file that js-yaml rejects with 'bad indentation
of a mapping entry', silently dropping custom_providers and breaking
model switching.
Add IndentDumper SafeDumper subclass that forces indentless=False,
route atomic_yaml_write through it. Route tui_gateway._save_cfg and
the Telegram adapter's config writer through atomic_yaml_write so all
paths emit the same 2-indent layout.
Salvaged from #32034 by @xxxigm. Adapted to current main which already
has allow_unicode=True (from #51356) but was missing IndentDumper.
Closes#31999
A prompt sent while a turn was in flight got rejected with 4009 "session busy",
which pushed clients (the desktop app) into a deadline-bounded busy-retry. When
turn teardown outlived that deadline — e.g. the user hits stop while a slow,
non-interruptible tool (web_search, read_file, an MCP call) is mid-flight, since
the sequential executor only checks the interrupt flag between tools — the
resubmitted message was silently dropped: "it just doesn't listen".
Wire the previously-dead display.busy_input_mode config into prompt.submit:
instead of rejecting, apply the policy and queue the message to run as the next
turn (drained in run()'s tail, ahead of goal/notification follow-ups). Modes:
interrupt (default) interrupts the live turn so it winds down promptly then runs
the queued message; queue runs it after the current turn finishes; steer injects
it into the live turn when accepted, else queues. The queued slot pins the
sender's transport and losslessly merges a second arrival. No client deadline,
no dropped sends.
The Desktop GUI (tui_gateway) slash worker subprocess has no reader for
the CLI's _pending_input queue. /learn's CLI handler prints the ack and
puts the built prompt onto that queue, so in the TUI the prompt was
silently dropped — ack shown, no LLM turn, no skill created (#51829).
command.dispatch already handles 'learn' correctly (returns
{type: send, message: build_learn_prompt(arg)}), but 'learn' was missing
from _PENDING_INPUT_COMMANDS, so slash.exec fell through to the worker
instead of routing to command.dispatch. Add it to the frozenset, matching
the existing goal/queue/steer/plan pattern.
Ship the final pet-generation UX polish (provider picker behavior, step-2 cancel flow, banner integration, and visual consistency) and make saturated-chroma background removal C-op driven so hatch processing no longer hammers the machine during long runs.
When no reference-capable image backend is configured, generating a pet is
impossible — so instead of a dead prompt + post-hoc error, the overlay now
detects it up front and offers a way out:
- pet.generate.status RPC reports whether a reference-capable provider
(OpenRouter / Nous Portal / OpenAI) is set up; the overlay probes it on
open and swaps the prompt for a friendly setup card (paw, one-line copy,
"Set up image generation" → /settings?tab=providers, key links).
- useRouteOverlayActive(): reusable hook so any portaled modal yields the
screen to a full-screen route overlay (e.g. settings) and reappears —
re-running its mount effects — on return, instead of closing. The probe
re-runs on that remount, so adding a key flips the card to the prompt.
- pet.generate / pet.hatch (parallel rows, off the reader thread) +
cooperative pet.cancel; pet.export / pet.rename.
- pet.gallery localOnly fast path + background manifest prefetch so the
picker never blocks on petdex; rename follows the active-pet config.
- gateway request gains optional timeout + AbortSignal for real Stop.
Let setup.runtime_check accept an optional provider, persist the selected
provider/model before the gate, and validate the provider the user just
connected instead of a stale config entry such as anthropic.
The TUI model-switch persistence (_persist_model_switch) rewrote the entire
model config block via save_config(), destroying sibling keys the user set
under model: (model_slots, model_fallback, base_url, ...) on every switch.
Use targeted, atomic, comment-preserving save_config_value("model.default" /
"model.provider" / "model.base_url") writes instead, so a model switch only
touches the keys it changes.
Salvaged from #48391 by kyssta-exe (authorship preserved).
Fixes#48305
atomic_yaml_write (and two sibling config writers) called yaml.dump
without allow_unicode=True. The default personalities shipped in cli.py
contain emoji/kaomoji, so PyYAML escaped astral-plane chars as 8-digit
\\UXXXXXXXX sequences inside multi-line double-quoted strings wrapped
with \\ line-continuations. Stricter/non-PyYAML parsers, editors, and
hand-edits break that structure into unclosed quotes, failing the whole
config parse -> silent fallback to defaults -> custom_providers lost.
Add allow_unicode=True to the canonical writer plus tui_gateway/server.py
and the telegram adapter's atomic config write so config is written as
readable UTF-8 with no escape/fold artifacts.
Fixes#51356
Open-ended skill learning across every surface. /learn <free text> takes a
description of any source — a directory, a URL, the workflow you just walked
the agent through, or pasted notes — and the live agent gathers it with the
tools it already has (read_file/search_files, web_extract, the conversation,
the pasted text), then authors a SKILL.md via skill_manage following the
house authoring standards (<=60-char description, the standard section order,
Hermes-tool framing, no invented commands).
No engine, no model-tool footprint, works on any terminal backend (local,
Docker, remote): /learn builds a standards-guided prompt and hands it to the
agent as a normal turn.
- agent/learn_prompt.py: shared standards-guided prompt builder
- /learn registry entry (both surfaces) + CLI handler (inject onto input
queue) + gateway handler (rewrite turn, fall through, /blueprint pattern)
- tui_gateway command.dispatch returns a send directive -> TUI + dashboard chat
- dashboard Skills page 'Learn a skill' panel (dir + URL + open-ended text)
composes a /learn request and runs it in chat
- docs (slash-commands ref + skills feature page), 11 targeted tests
Inspired by OpenAI Codex's Record & Replay and the /learn concept from #47234
(dir-distillation engine); reworked to be open-ended and engine-free per
review.
When a session rotates id on compression, _sync_session_key_after_compress()
re-anchored the session_key, approval-notify routing, yolo state, and slash
worker — but never moved the active-session lease, which stayed keyed to the
pre-compression id. And _find_live_session_by_key() matched live sessions on
the stale session_key, not the live agent's current agent.session_id. After
compression a resume/create path failed to recognize the existing live agent
and could build a SECOND live agent against the same DB continuation -> forked
lineage / cross-session message mixing.
- active_sessions.transfer_active_session(): move a lease in place to the new
id under the exclusive file lock (no slot drop).
- gateway _transfer_active_session_slot(): call it inside
_sync_session_key_after_compress(); on the rare fallback (entry pruned)
RESERVE the new slot before releasing the old lease (reserve-before-release),
so a concurrent gateway at the session cap cannot grab the freed slot in a
release-then-reacquire window and leave this session with no lease; if the
reserve fails, keep the existing lease (review fix).
- _session_lookup_key(): make live-session lookup authoritative on
agent.session_id, wired into all stale-session_key consumers
(_find_live_session_by_key, _session_live_item, _live_session_payload) —
fixes the whole lookup class.
Co-authored-by: kshitijk4poor <82637225+kshitijk4poor@users.noreply.github.com>
Parity with the classic CLI status bar's ⛓ indicator (PR #51441). The
Ink TUI status bar now shows ⛓ N for live background/async subagents
(delegate_task batches + background single delegations).
- tui_gateway/server.py: _get_usage() embeds active_subagents from
tools.async_delegation.active_count() — the same registry the CLI
reads — onto the existing per-update usage payload, guarded so a
raising active_count() leaves the field off without breaking usage.
- ui-tui appChrome: new 'subagents' status segment (breakpoint w>=92,
slots between bg and cost in the shed-order), renders ⛓ N from
usage.active_subagents.
- Usage / SessionUsageResponse types gain active_subagents?.
Distinct from the turn-scoped SpawnHud / /agents overlay, which mirror
live in-turn subagent.* events; this is the persistent registry count.
A "one-shot" is a single stateless model call that runs OUTSIDE any conversation:
it never touches session history, never breaks prompt caching, and returns plain
text. UI surfaces need this for small generative chores — a commit message from a
diff, a rename suggestion, a summary — where an agent turn would pollute the
thread and hand-rolling an LLM call at every call site would be worse.
- `agent/oneshot.py`: `run_oneshot(...)` over the existing auxiliary-client
plumbing (same path as title generation). Two call shapes: explicit
instructions/input, or a registered `template` + `variables` (templates own the
prompt engineering so it stays consistent across CLI/TUI/desktop). Ships a
`commit_message` template. Model selection inherits the live session via
`main_runtime`, else the configured aux `task` backend.
- `tui_gateway/server.py`: `llm.oneshot` RPC (long-handler) inheriting the
session's model when `session_id` resolves.
Stateless by construction — no session mutation, cache untouched.
Follow-up to the coding-context posture (#43316): that PR detects each repo's
verify loop (manifests, package manager, exact test/lint/build commands, context
files) and bakes it into the system-prompt snapshot — but only as a string, for
the model. Non-prompt consumers (the desktop verify UI) had no way to read it
without re-sniffing and drifting from the prompt.
Split detection from rendering, keeping one source of truth:
- `detect_project_facts(root) -> ProjectFacts` (frozen) holds the structured
facts; `_project_facts()` now renders it into the same snapshot lines, so the
prompt block stays byte-identical (cache-safe).
- `project_facts_for(cwd)` resolves the workspace root (git, else marker) and
returns the structured facts, or None outside a workspace.
- `project.facts` gateway RPC surfaces it to any client (desktop/TUI/ACP).
Tests assert the structured output and that the UI-facing commands never drift
from what the prompt block renders (one detector feeds both).
Follow-up to #50767, which redacted the chat-platform (_approval_notify_sync)
and SSE/API (_approval_notify) approval transports. The TUI JSON-RPC transport
is the third egress and was missed: three register_gateway_notify callbacks in
tui_gateway/server.py emitted the raw approval_data — including the unredacted
command Tirith flagged — straight to the TUI client via _emit.
Route all three registrations through a new module-level _emit_approval_request()
helper that redacts payload['command'] via the shared
gateway.run._redact_approval_command seam before emitting, matching the pattern
used for the other two transports. Completes the whole-bug-class fix for #48456.
Tests: assert the helper emits a redacted command (real credential pattern),
handles missing/None command, and a wiring guard that no registration emits the
raw payload directly (only the helper may). Both mutation-checked.
The #48456 fix series originated from @liuhao1024's #48462 — credit to them for
the original report and chat-platform fix; this completes the remaining transport.
Co-authored-by: liuhao1024 <sunsky.lau@gmail.com>
* feat(goals): add /goal wait <pid> barrier to park the loop on a background process
The /goal loop re-pokes the agent every turn via the post-turn judge. When a
goal is gated on a long-running background process (CI poller, build, test
matrix, deploy) that produces nothing to judge yet, this spins the agent into
'is it done?' busy-work and burns the turn budget.
/goal wait <pid> [reason] parks the loop: while the PID is alive, the judge is
skipped, no turn is consumed, no continuation fires, and /goal status shows a
parked indicator. The barrier auto-clears the moment the process exits (the
agent's notify_on_complete watcher is the natural wake signal), then the next
turn resumes normal judging. /goal unwait clears it manually; pause/resume/clear
drop it; a dead/stale PID can never wedge the loop.
Wired across CLI, gateway, and the mid-run command guard for parity. Barrier
persists in SessionDB.state_meta (survives /resume); GoalState gains
backward-compatible waiting_on_pid/waiting_reason/waiting_since fields. 12 new
tests; docs updated.
* fix(goals): use gateway.status._pid_exists for liveness, not os.kill(pid,0)
The Windows-footguns CI guard flagged os.kill(pid, 0) in _pid_alive — on
Windows that's not a no-op, it routes to CTRL_C_EVENT and hard-kills the
target's console process group (bpo-14484). Delegate to the canonical
footgun-safe gateway.status._pid_exists (psutil + ctypes/POSIX fallback)
instead, with a direct-psutil last resort.
* feat(goals): judge-driven auto-wait — the loop parks itself, no manual /goal wait
Makes the wait barrier automatic. Every turn the judge is shown the agent's
live background processes (pid, command, uptime, output tail from the
process_registry) alongside the goal + response, and can return a new 'wait'
verdict instead of continue:
{"verdict":"wait","wait_on_pid":N} → park until that process exits
{"verdict":"wait","wait_for_seconds":N} → park until the deadline passes
evaluate_after_turn acts on the directive (sets the barrier, parks the loop)
so the agent isn't re-poked into busy-work while CI/builds/deploys run. Adds a
time-based waiting_until barrier alongside the pid barrier; both auto-clear and
can never wedge the loop. Drivers (CLI, gateway, tui_gateway) feed the live
registry in via gather_background_processes(). Manual /goal wait stays as an
override. Judge verdict contract widened to (verdict, reason, parse_failed,
wait_directive); legacy {"done":bool} shape still accepted.
* test(goals): update kanban _fake_judge to the 4-tuple judge contract
CI test(3) caught it: test_kanban_goal_mode's _fake_judge still returned the
3-tuple (verdict, reason, parse_failed), but the kanban loop now unpacks the
4-tuple (+ wait_directive). Update the fake to return None for the directive
and accept the background_processes kwarg.
* feat(goals): trigger-based wait — park on a process's own signal, not just exit
Addresses two gaps in the judge-driven wait: (1) the judge could only express
'wait until PID exits' or 'wait N seconds', so a long-lived watcher/server that
fires a trigger MID-RUN (and may never exit) couldn't be waited on; (2) the
process's own watch_patterns/notify_on_complete trigger was invisible to the judge.
Adds a session-based barrier (waiting_on_session) that releases on the process's
OWN trigger via process_registry.is_session_waiting(): the session exits, OR (if
started with watch_patterns) its pattern matches — even while the process keeps
running. list_sessions() now surfaces session_id + watch_patterns/watch_hit/
notify_on_complete so the judge sees the trigger and is told to prefer
wait_on_session for trigger processes. Judge verdict gains a {wait_on_session}
directive (preferred over pid). Backward-compatible GoalState field; pid + time
barriers unchanged.
Tests: TestSessionTriggerBarrier (release on mid-run pattern match while alive,
release on exit, unknown-session, full park→trigger→resume, parse, validation,
backcompat load). 105 goal-surface + 85 process_registry tests green.
* feat(cli): /reasoning full to show complete thinking, not 10-line clamp
The post-response Reasoning recap box hard-clamped long thinking to the
first 10 lines, so there was no way to see the full reasoning trace after
a turn (live streaming already shows it in full). Add display.reasoning_full
(default off) plus /reasoning full|clamp to toggle it at runtime; the clamp
truncation note now points at the command. Addresses repeated user requests
to show all thinking tokens.
* test(gateway): de-snapshot /reasoning help assertion
The test froze the exact args-hint literal '/reasoning [level|show|hide]',
which the new full/clamp args change to '[level|show|hide|full|clamp]'.
Convert to an invariant: assert /reasoning is in help and carries its core
args, not the exact hint string.
* feat(tui): /reasoning full|clamp parity in tui_gateway
The classic-CLI reasoning_full toggle had no TUI equivalent — typing
/reasoning full in the TUI fell through to parse_reasoning_effort and
errored. The TUI renders thinking as an expand/collapse section (no fixed
10-line recap), so map full -> sections.thinking=expanded (raw, uncapped
via thinkingPreview mode='full') and clamp -> collapsed, persisting
display.reasoning_full for cross-surface config consistency.
When a /model switch resolves a valid model but the in-place agent swap
fails mid-conversation (expired key, unreachable base_url), the agent
rolls itself back to the old working model+client and re-raises. The
callers caught that re-raise, logged a warning, then committed the broken
switch anyway: wrote the failed model to the session DB, set
_session_model_overrides to the broken model/provider/key, and (gateway
direct path) evicted the working cached agent. The next message then
rebuilt a dead agent from the broken override -> permanently unusable
conversation (#50163).
Fix the whole caller class so a failed swap aborts the commit entirely:
- gateway/slash_commands.py (picker + direct /model paths): on swap
failure, early-return an error message; skip DB persist, session
override, cache eviction, and config write.
- cli.py (both /model handlers): snapshot CLI-level credential/runtime
fields before mutating, restore them on swap failure, and abort the
note + success print.
- tui_gateway/server.py: wrap the previously-unguarded swap; on failure
raise a clean error and skip worker restart, runtime persist, switch
marker, session model_override, and config persist.
The no-cached-agent path (apply-on-next-session) is unaffected.
Adds a gateway regression test that fails on the pre-fix behavior.
The TUI /compress slash side-effect compressed the session, synced the
key, and emitted session.info — but returned an empty string, so the
user saw no 'Compressed: N → M messages / ~X → ~Y tokens' feedback. The
CLI (_manual_compress) and gateway (slash_commands) paths both already
call summarize_manual_compression; the TUI slash path was the lone gap.
Snapshot history + rough token estimate before and after compaction and
return the formatted summarize_manual_compression() feedback, mirroring
the session.compress RPC handler. The estimate uses the same
estimate_request_tokens_rough(system_prompt, tools) inputs as the RPC
path, re-reading the system prompt after compaction (it may be rebuilt).
Co-authored-by: liuhao1024 <sunsky.lau@gmail.com>
Mirror the CLI's exit-path behaviour in the TUI gateway so that
unpersisted conversation messages are flushed to state.db and the
on_session_end plugin hook fires before the session is closed.
Root cause: _finalize_session() only called db.end_session() to
mark the session row as ended, but did NOT flush in-memory messages
via _persist_session() or fire the on_session_end hook. When the
user force-quit (double Ctrl-C, terminal-close, SIGHUP) while the
agent was mid-turn, messages accumulated since the last persist
point were silently lost.
Changes
-------
tui_gateway/server.py - _finalize_session():
- Persist unflushed messages via agent._persist_session() before
db.end_session(). Prefers agent._session_messages (set by the
last _persist_session call inside run_conversation) over
session['history'] (stale when agent is mid-turn).
- Fire on_session_end(interrupted=True) plugin hook so crash-
recovery plugins can flush buffers, matching cli.py behaviour.
tui_gateway/entry.py - _log_signal():
- Explicitly call _shutdown_sessions() before sys.exit(0) in the
SIGHUP/SIGTERM handler as belt-and-suspenders over atexit.
tests/tui_gateway/test_finalize_session_persist.py (new):
- 11 tests covering: history persistence, _session_messages
priority, empty-history skip, missing-agent, double-finalize,
persist-exception resilience, hook firing, hook-exception
resilience, and db.end_session preservation.
Related
-------
Closes the TUI half of #5021 (CLI already handles this via its
atexit handler). Also addresses the session-persistence gap
discussed in #18465 and #18269.
Follow-up to the salvaged preflight-compression warning:
- Replace silent `except Exception: pass` at all 5 guard call sites
(cli.py x2, gateway/slash_commands.py x2, tui_gateway/server.py) with
`logger.debug(...)` so signature drift in the guard helper isn't hidden.
- tui_gateway/server.py: set the confirm dict's `warning` field to the
merged message (was bare expensive-model text) so it matches
`confirm_message` for any future consumer reading `warning`.
- Add trailing newlines to the two new files.
Adds hermes_cli/context_switch_guard.py mirroring the model_cost_guard
pattern. When a user switches models mid-session (Herm TUI picker, CLI,
or /model on Telegram/Discord), the warning surfaces on the existing
ModelSwitchResult.warning_message path used by the expensive-model
guard if the new model's compression threshold is below the current
session size.
Partial fix for #23767 — addresses only the 'user-facing guardrail
when switching from a high-context provider to a substantially
lower-context provider' slice. The other proposed fixes from that
issue (hard preflight token guard, metadata cache invalidation on
switch, compression safety invariant, oversized tool-output handling)
are out of scope for this PR.
The MCP discovery wait is now bounded by the config-driven mcp_discovery_timeout
(default 1.5s), not the old 0.75s flat value. Updates the _schedule_mcp_late_refresh
docstring that still cited ~0.75s after #49208 made the bound configurable.
Add the Ink pet sprite pane, the interactive /pet picker overlay, and live
pet switching/rescale driven by new tui_gateway RPCs (pet state, pet.scale,
per-state frames). Wires pet flash state and the picker into the TUI layout
and slash handler. Covered by the slash-handler test.
Consolidated findings from three independent reviewers (Codex, Claude Code, a
Hermes subagent w/ the hermes-agent-dev skill):
- BLOCKING: refresh_agent_mcp_tools rebuilt only the registry subset, silently
dropping post-build-injected memory-provider (mem0/honcho/…) and context-
engine (lcm_*) tools on every refresh. Now additive-preserving: re-applies
the same injectors agent_init uses, staged on locals and published atomically.
- Re-injection now honors the #5544 enabled_toolsets gate for context-engine
tools, so a restricted-toolset platform can't get lcm_* leaked back in.
- Atomic read-diff-publish under one lock: the returned `added` set and the
(tools, valid_tool_names) pair are consistent even under concurrent callers
(no half-swap, no TOCTOU).
- background_review fork opts out (_skip_mcp_refresh) so its byte-identical
tools[] cache parity with the parent is preserved.
- CLI /reload-mcp routed through the shared helper (was a 4th divergent copy
with the same clobber bug + missing disabled_toolsets).
- Explicit reloads (TUI RPC + CLI) pass enabled_override so a server the user
just enabled in config this session is picked up; automatic paths reuse the
agent's build-time selection.
- mcp_discovery_timeout default 5.0 -> 1.5s: correctness now comes from the
between-turns refresh, so the startup wait is only a small turn-1 UX bump
rather than a heavy dead-server latency penalty.
- has_registered_mcp_tools checks registered TOOLS (not connected servers) so a
zero-tool/prompt-only server doesn't make the per-turn hook fire forever.
- Tests: rewrote the thread-safety test to actually exercise the write path
(alternating tool sets), added the #5544-gate regression, the memory/context
preservation regression, and a "callable next turn via valid_tool_names"
contract; removed a dead monkeypatch line.
MCP servers that connect after the agent's one-time tool snapshot were
invisible for the whole session. Two root causes, fixed together:
1. The startup discovery wait was a flat 0.75s. HTTP/OAuth servers
commonly take 2-6s on a cold connect, so they missed the window and
their tools never entered the agent's snapshot. `thread.join(timeout)`
already returns the instant discovery completes, so raising the bound
costs ~0s for the common case (no MCP / fast servers) and only ever
blocks for a genuinely-pending server, capped so a dead server can't
freeze startup. The bound is now configurable via
`mcp_discovery_timeout` (config.yaml, default 5.0s).
2. Three call sites duplicated the agent tool-snapshot rebuild (the TUI
`reload.mcp` RPC, the gateway reload, and the TUI late-binding refresh
thread), and the late-refresh detected changes by tool COUNT — missing
an equal-size add/remove swap. Consolidated into one shared
`tools.mcp_tool.refresh_agent_mcp_tools(agent)` helper that diffs by
tool NAME, mutates the agent under a lock (thread-safe), and respects
the agent's own enabled/disabled toolsets.
The late-binding refresh keeps its pre-first-turn cache-safety guard:
it never rebuilds the tool list once a turn has started, so the cached
prompt prefix is never invalidated mid-conversation.
Tests: new tests/tools/test_refresh_agent_mcp_tools.py covers the
name-based diff, in-place mutation, agent-scoped filtering, thread
safety, and the config-driven discovery bound (incl. instant-return
when nothing is pending). 75 passed across the touched areas.
The skin bug was one instance of a class: several subsystems build their
config dict directly from config.yaml instead of routing through
hermes_cli.config.load_config (which carries the managed merge), so they
silently ignored administrator-pinned values. Audited every config.yaml
reader and fixed the behavioral-read bypasses:
- gateway/config.py load_gateway_config (messaging gateway: session_reset,
quick_commands, stt, model, ...)
- gateway/run.py _load_gateway_config (its read_raw_config fast path also
skipped the merge — read_raw_config returns raw user YAML)
- tui_gateway/server.py _load_cfg (new TUI + desktop backend: skin,
reasoning_effort, service_tier, provider_routing)
- cron/scheduler.py (scheduled-job model/reasoning/toolsets/provider_routing)
- hermes_logging.py (logging.level/max_size_mb/backup_count)
- hermes_time.py (timezone)
- hermes_cli/doctor.py (memory-provider diagnostic reads effective config)
All route through a new shared managed_scope.apply_managed_overlay() helper
that mirrors _load_config_impl (env-only expansion so a user ${VAR} can't
shadow a managed literal, root-model-string normalization, leaf-merge) and is
fail-open. cli.py's earlier inline fix is refactored onto the same helper.
Write-back paths (slash_commands, telegram/yuanbao dm_topics, profile
distribution) are deliberately left reading raw user YAML — overlaying managed
values there would persist them into the user file. The dashboard
(web_server.py) already routes through load_config and needed no change.
TUI loader caches the RAW config so _save_cfg never writes managed values to
disk. Adds test_managed_scope_overlay.py (helper) and
test_managed_scope_loaders.py (per-surface integration); mutation-checked.
A plain /model <name> switch only lasted for the current session — every
new session reverted to the previously-configured model, so users had to
re-switch every time (e.g. glm-5.1 -> glm-5.2 on every launch).
Persist-by-default is now the behavior across all three /model surfaces
(CLI, gateway, TUI/dashboard), gated by a new config key
model.persist_switch_by_default (default true):
/model <name> switch model (persists to config.yaml)
/model <name> --session switch for this session only
/model <name> --global switch and persist (explicit, unchanged)
The effective persistence is resolved once via resolve_persist_behavior()
in hermes_cli/model_switch.py so --session opts out, --global opts in,
and the config-gated default applies otherwise. --global remains a valid
explicit no-op alias for the new default.
When /goal (and other _PENDING_INPUT_COMMANDS: retry, queue, q, steer,
plan, undo) were typed in the TUI desktop app, slash.exec returned error
4018 instructing the frontend to fall back to command.dispatch. Some
clients failed that client-side fallback, leaving the command empty and
surfacing "empty command" — the user's typed text was silently dropped.
slash.exec now routes pending-input commands to command.dispatch
internally, eliminating the fragile client-side fallback hop. The
response is exactly what command.dispatch would have produced, so the
TUI client behaves identically once the round-trip succeeds.
Salvaged from #48944 — rebased onto current main. The original PR's
source change and test_goal_command.py update are correct, but it missed
the second test surface: tests/tui_gateway/test_protocol.py's
parametrized test_slash_exec_rejects_pending_input_commands still
asserted the old 4018 rejection for retry/queue/q/steer/plan, turning CI
red (5 failures). That test is rewritten here as a behavior contract:
slash.exec for a pending-input command must yield the same payload as a
direct command.dispatch call, and must no longer emit the old
"pending-input command" fallback rejection.
Co-authored-by: kyssta-exe <kyssta-exe@users.noreply.github.com>
The desktop model picker had no way to force a fresh model fetch: model.options
went through the 1h-cached provider_models_cache.json, and there was no flag to
bust it. When a provider's cached list expired and its next live fetch failed,
the picker fell back to the curated static list — silently dropping live-only
models (e.g. OpenCode Zen's free tier like deepseek-v4-flash-free) the user had
been using.
- Thread refresh through model.options (RPC + REST /api/model/options) ->
build_models_payload -> list_authenticated_providers, which calls
clear_provider_models_cache() up front when set so every row re-fetches live.
- Add a 'Refresh Models' control to the desktop picker (5-locale i18n, spinning
sync icon). Normal opens leave refresh=false to stay snappy on the cache.
Verified: stale cache hides deepseek-v4-flash-free -> refresh busts it -> live
re-fetch surfaces it. refresh=false never touches the cache.
Auto-compression ends the live session and forks a continuation child
(linked via parent_session_id). A long-lived parent keeps its own flushed
message rows, so resolve_resume_session_id()'s empty-head walk never
redirected it — resuming the parent id reloaded the pre-compression
transcript and dropped every turn generated after compression, including
the assistant's response. On the desktop this is the recurring "I sent a
message, came back, and the reply isn't there" report on large sessions:
the chat's routed id is the pre-rotation id, and both the gateway
session.resume RPC and the REST /messages read anchored on it.
Fix the resolver at the chokepoint: resolve_resume_session_id() now
follows the compression-continuation chain forward via get_compression_tip()
before its existing empty-head descendant walk. get_compression_tip() only
follows children whose parent ended with end_reason='compression' (created
after the parent was ended), so delegation/branch children never hijack a
resume. This fixes every resume caller at once (REST /messages, CLI
--resume, gateway /resume).
session.resume in tui_gateway was the one resume path that never called the
resolver — it used the raw target id directly. Route it through
resolve_resume_session_id() too (non-lazy only; lazy watch windows must
stay on their exact child branch). Resolving up front also re-anchors the
live-session fast path so a still-live rotated session is reused by its new
key instead of rebuilding a duplicate agent on the stale parent.
Tests:
- resolve_resume_session_id follows the tip even when the parent retains
messages, and is not confused by a delegation child.
- session.resume binds the agent to the continuation tip and returns the
post-compression reply.
The interactive model pickers (Desktop REST API, TUI model.options, CLI
/model) were hard-capped at max_models=50, which truncated large provider
catalogs like Kilo Gateway (336 models) to just 50 entries. This made
most models undiscoverable via the picker search box.
Changes:
- Change build_models_payload() default from max_models=50 to None (unlimited)
- Change list_authenticated_providers() default from max_models=8 to None
- Change list_picker_providers() default from max_models=8 to None
- Fix all [:max_models] slicing to handle None as 'no limit'
- Remove max_models=50 from 5 interactive picker callers:
* web_server.py: get_model_options (Desktop /api/model/options)
* web_server.py: get_recommended_default_model
* model_switch.py: prewarm_picker_cache_async
* tui_gateway/server.py: model.options JSON-RPC
* cli.py: HermesCLI model picker
- Telegram/Discord inline keyboard picker (gateway/slash_commands.py)
still passes max_models=50 explicitly — unchanged behavior.
The total_models field was already in the response payload and is now
meaningful since models.length == total_models for interactive pickers.
Fixes#48279
* feat(billing): nous_billing http client + BillingState core (phase 2b)
Phase 2b terminal-billing client foundation:
- hermes_cli/nous_billing.py: typed client for the 4 /api/billing/* endpoints
(state/charge/poll/auto-top-up). Raises typed errors (BillingScopeRequired,
BillingRateLimited, BillingAuthError) mapped from the live-verified contract;
fail-open is the caller's job. Idempotency-Key enforced client-side.
- agent/billing_view.py: surface-agnostic BillingState core + Decimal money
parsing (server emits decimal strings, not 2dp), fail-open builder,
idempotency-key gen, custom-amount validation.
- 51 unit tests (decimal parse/format, payload tiering, error->exception
matrix, fail-open, amount validation).
Plan: docs/plans/2026-06-13-001-phase-2b-terminal-billing-tui-plan.md
* feat(billing): billing:manage scope + lazy step-up re-auth (phase 2b)
- NOUS_BILLING_MANAGE_SCOPE constant.
- nous_token_has_billing_scope(): split-based scope check (no false-positive
substring match).
- step_up_nous_billing_scope(): re-runs the device flow requesting
billing:manage, reusing the held credential's portal/inference URLs + client_id
(so a preview stays a preview), persists like _login_nous but WITHOUT the model
picker. Returns True iff the minted token carries the scope (False when NAS
silently downscopes a non-admin / unticked grant).
Lazy step-up (plan D-A): normal login path unchanged; 403 insufficient_scope
from a billing call triggers this. 7 unit tests.
* feat(billing): billing JSON-RPC methods for the TUI (phase 2b)
billing.state / charge / charge_status / auto_reload / step_up in
tui_gateway/server.py. Return STRUCTURED success envelopes (result.ok +
result.error=<code>) rather than JSON-RPC-level errors, so the Ink rpc() promise
always resolves and the TUI branches on the typed billing error code
(insufficient_scope, rate_limited, no_payment_method, …) to render the right
affordance. Money serialized as decimal STRINGS + display strings. charge mints
+ echoes an idempotency_key for retry reuse. 16 unit tests.
* feat(billing): /billing CLI handler + command registry (phase 2b)
- CommandDef("billing", subcommands=buy|auto-reload|limit), added to
_SLACK_VIA_HERMES_ONLY so it routes via /hermes on Slack (keeps the 50-cap
parity test green, same as /credits).
- cli.py::_show_billing + screen helpers: all 5 screens (overview, buy→confirm→
poll, auto-reload, monthly-limit read-only). Reuses _prompt_text_input_modal /
_prompt_text_input (D-C). Non-interactive (_app is None) renders text + portal
deep-link, never prompts (R7). Decimal money end-to-end. 2s/5-min cancellable
poll loop; 429/503 = retry not failure; settled = ledger truth. Lazy step-up on
403 insufficient_scope. no_payment_method treated as mainline funnel-to-portal.
- 6 CLI tests; 156 command tests (incl. Slack/Telegram parity) green.
* feat(billing): /billing Ink TUI screens + tests (phase 2b)
- ui-tui/src/app/slash/commands/billing.ts: /billing TUI command covering all 5
screens — overview (text), buy <amt> → ConfirmReq → charge → non-blocking 2s/
5-min poll loop → settled/failed/timeout branches, auto-reload <below> <to> →
ConfirmReq → PATCH, limit (read-only). Reuses the existing ConfirmReq overlay
(D-C) — no bespoke component. Typed-error envelope branching: insufficient_scope
arms the lazy step-up confirm; no_payment_method/rate_limited/cap funnel to
portal. Client-side amount validation mirrors the server (bounds + 2dp).
- gatewayTypes.ts: Billing* response interfaces.
- registry.ts: register billingCommands.
- billingCommand.test.ts: 12 vitest cases (overview/gating/buy-confirm-poll-
settled/no_payment_method/step-up/limit/auto-reload/validation).
TUI build green; 12/12 vitest pass; slash tests pass once @hermes/ink is built.
* docs(billing): scrub private cross-repo references
NAS is a private repo — remove all references to it from the public PR:
- drop the cross-repo planning doc (planning scaffolding, not a deliverable;
the PR description documents the design)
- replace 'NAS' / 'PR #412 preview' mentions in code + test comments with
generic 'the server' / 'a preview deployment'
* docs(billing): scrub final NAS reference in step-up docstring
* docs(billing): drop dangling plan-doc refs
The phase-2b plan doc was removed in the cross-repo scrub (300afcc0b)
but two module docstrings still pointed at it. Drop the dead refs.
* feat(billing): interactive /billing overlay + step-up UX, portal-URL & token fixes
Adds the interactive /billing TUI overlay and hardens the terminal-billing
client across CLI and TUI.
- TUI: full /billing overlay state machine (overview to buy to confirm,
auto-reload, read-only monthly limit) reusing the existing confirm overlay.
- Step-up: surface the verification link in-transcript and open the browser
via the TUI's own opener (the device flow runs in the headless gateway, so a
printed URL was being dropped); run the step-up handler off the main loop and
emit the link as an out-of-band event so the gateway stays responsive.
- Step-up copy is scope-accurate ("Billing permission granted") and re-checks
/state so it never claims "enabled" when the org kill-switch is still off.
- Portal deep-links resolve to absolute URLs against the active portal base
(the server emits them relative) - fixes a bare "/billing?topup=open" link.
- Billing calls refresh an expired access token via the stored refresh token
instead of reporting a false "not logged in".
- Optimistic funnel: advise "set up a saved card on the portal" up front when
no card is on file (advisory, not a hard gate).
- Token resolution is cached briefly so the 2s charge poll loop stops
re-locking + re-reading the auth store on every tick; 401 re-resolves fresh.
- Remove the temporary demo-mode shims.
Validation: 87 Python billing tests, 88 TS tests (billing command + gateway
event handler), tsc clean, ink + ui-tui builds green.
* docs(billing): add /billing TUI screenshots for PR
* fix(cli): guard _last_invalidate on bare instances; update stale prompt-fallback test
The UI-invalidate throttle read self._last_invalidate unconditionally, which
raised AttributeError on HermesCLI instances built without __init__ (the
thread-safety test's object.__new__ shell). Guard the read with getattr.
The off-main-thread branch of _prompt_text_input was changed (#23185) to cancel
cleanly to None instead of falling back to a bare input() that would hang on the
slash-worker thread; the test still asserted the old direct-input fallback.
Update it to assert the current intended behavior: returns None, calls neither
run_in_terminal nor input(), and does not hang.
The "💾 Self-improvement review" summary (skill/memory updated) was invisible
on two surfaces:
- Desktop Electron app had no review.summary event handler — skill/memory
writes happened silently. Now appends a persistent system message to the
transcript (matching the Ink TUI's persistent-line semantics, not a
transient toast that can be missed).
- tui_gateway (backs both 'hermes --tui' and the desktop) never read
display.memory_notifications, so it always behaved as 'on' and ignored a
user who set 'off'/'verbose'. Added _load_memory_notifications() (mirrors
the messaging gateway's bool->str normalization, defaults to 'on') and
wired it to agent.memory_notifications, matching gateway/run.py and the CLI.
Delivery chain now reaches all surfaces:
background_review.py -> background_review_callback -> review.summary event ->
desktop transcript / Ink TUI line / gateway message / CLI print.
* Port from cline/cline#11514: encourage parallel tool calls
Add a universal system-prompt guidance block telling the model to batch
independent tool calls (reads, searches, web fetches, read-only commands)
into a single assistant turn instead of one call per turn. The runtime
already executes independent batches concurrently (read-only tools always;
non-overlapping path-scoped file ops); the open-source system prompt had
nothing steering the model to PRODUCE the batch. Fewer round-trips means
less resent context, which compounds over a long conversation.
- prompt_builder.py: new PARALLEL_TOOL_CALL_GUIDANCE block (short, static,
cache-amortised) modeled on TASK_COMPLETION_GUIDANCE.
- system_prompt.py: inject right after the task-completion block, gated by
agent.valid_tool_names + the new toggle.
- agent_init.py: read agent.parallel_tool_call_guidance (default True).
- config.py: add the default under the agent section.
- test_prompt_builder.py: behavior-contract tests (batching steer, dependent
carve-out, length bound) — invariants, not wording snapshots.
Adapted from Cline's TypeScript tool-surface guidance to hermes-agent's
Python prompt-assembly architecture and config-over-env conventions.
* fix(desktop): never persist or restore a named custom provider as bare "custom"
Custom providers vanish from the Desktop/TUI model picker with
"No LLM provider configured" — repeatedly fixed (#44062, #44109, #45578)
and repeatedly regressed (#44022, #47714) because every fix only recovered
the entry identity from a persisted base_url. When a session is
persisted/restored with the resolved provider "custom" and NO base_url, bare
"custom" leaked through verbatim; resolve_runtime_provider("custom") routes to
the OpenRouter default URL with no api_key, so the next turn/resume dies.
Bare "custom" is the resolved billing class shared by every named providers:/
custom_providers: entry — it is not a routable identity. Centralize the
"never let bare custom escape" invariant in one helper,
runtime_provider.canonical_custom_identity(), and apply it at all four leak
sites in tui_gateway/server.py:
- _ensure_session_db_row — the ORIGIN: first DB write seeds the bad row
- _runtime_model_config — live persist
- _stored_session_runtime_overrides — resume restore (heals old rows; drops
unrecoverable bare custom so resume falls back to config default)
- _make_agent — rebuild / per-turn
The helper recovers custom:<name> from the endpoint URL when present, else
from config.model.provider (the durable identity left when no base_url
survived). Regression tests in test_custom_provider_session_persistence.py
lock the no-base_url vector at every site so it cannot regress again.
