Commit Graph

3605 Commits

Author SHA1 Message Date
Teknium d48a0e604f Merge branch 'simp/r2-gw-rest-config' into simp/integration2 2026-09-02 17:05:22 -07:00
Teknium a3b6d91302 Merge branch 'simp/r2-update-doctor' into simp/integration2 2026-09-02 17:05:20 -07:00
Teknium 74b4d698fc Merge branch 'simp/r2-config-tools' into simp/integration2 2026-09-02 17:05:18 -07:00
Teknium d6f56549c0 Merge branch 'simp/r2-webserver' into simp/integration2 2026-09-02 17:05:16 -07:00
Teknium 93673653fa Merge branch 'simp/r2-kanban' into simp/integration2 2026-09-02 17:05:05 -07:00
Teknium cba6a6de8e Merge branch 'simp/r2-clipy' into simp/integration2 2026-09-02 17:05:04 -07:00
Teknium 5820d0b0d5 refactor(gateway/config): extract the config.yaml phase to gateway/config_loader.py; table-driven from_dict; defensive collapse
- load_gateway_config (536 LOC) is now a thin orchestrator: legacy gateway.json
  -> config_loader.load_yaml_layer -> GatewayConfig.from_dict -> env overrides
  -> validation. The yaml phase lives in gateway/config_loader.py as small
  functions driven by tables: _TOPLEVEL_BRIDGE (23 top-level/nested
  gateway.<key> bridges with 5 fallback modes), _SHARED_KEYS (28 per-platform
  keys copied into extra, with per-platform restrictions and transforms) and
  _PORT_BRIDGE_KEYS. Logger name kept as "gateway.config".
- GatewayConfig.from_dict: shared pick()/key_label() helpers replace the
  repeated "top-level key present else nested gateway.<key>" blocks; warning
  order preserved.
- _normalize_unauthorized_dm_behavior / _normalize_notice_delivery unified into
  _normalize_choice; _ensure_platform_extra_dict -> _dict_slot (also used by
  persist_home_channel); _getenv_int removed (dead since the env pass moved to
  config_env, which has its own _int_or).
- Platform._missing_: one _add_pseudo_member helper for both branches.
- Single warning sites in _coerce_optional_positive_int and
  coerce_systemd_watchdog_seconds; _validate_gateway_config placeholder pass
  flattened; small to_dict/getter collapses. Ruff F401/SIM102 clean.
- tests/hermes_cli/test_config_read_guard.py: allowlist gateway/config_loader.py
  (same owner as gateway/config.py — the extracted load_gateway_config phase).

gateway/config.py 2684 -> 1319 LOC (-50.9%); largest function now
GatewayConfig.from_dict at 119 LOC. Resolved-config parity: 160 cells
(16 yaml fixtures x 10 env sets) byte-identical to the integration base,
including captured log records and stderr.
2026-09-02 16:48:36 -07:00
Teknium e03a680592 refactor(kanban-db): split into connect/workspace/dispatch/notify modules
Byte-identical bodies moved out of hermes_cli/kanban_db.py into four sibling
modules, re-exported from the origin so kanban_db.<name> keeps resolving and
stays the single monkeypatch target; origin-resident helpers are reached via a
late-bound _kb namespace. AST-identity verified for all 329 moved symbols; SQL
statement multiset parity vs base. Three source-inspection tests repointed at
kanban_db_dispatch; the _add_column_if_missing alias test now imports the real
owner (hermes_cli.sqlite_util).
2026-09-02 16:46:52 -07:00
Teknium 79aeb22886 refactor(cli): unify duplicated model-switch staging and last-assistant-text blocks
cli_model_switch_mixin: the identical snapshot->stage->agent.switch_model->rollback
block in _apply_model_switch_result and _confirm_and_apply_cli_model_switch (2x47
lines) -> _stage_and_swap_model(result, old_model) -> bool. cli_loops_mixin: the
identical reversed-history assistant-text extraction in the loop-tick and goal
post-turn hooks -> _last_assistant_response_text(). Four _StubCLI test doubles bind
the new helper to the real implementation.
2026-09-02 16:17:31 -07:00
Teknium ffed5b1503 refactor(tools_config): split tools_command into flow helpers; drop unused imports 2026-09-02 16:14:49 -07:00
Teknium 13e7cb8eb9 refactor(tools_config): merge provider-selection extraction (tools_config_providers) 2026-09-02 16:11:56 -07:00
Teknium 3f0e1165b3 refactor(tools_config): extract post-setup hooks into tools_config_post_setup 2026-09-02 16:11:26 -07:00
Teknium c11fa53d92 refactor(tools_config): extract provider selection cluster into tools_config_providers 2026-09-02 16:05:46 -07:00
Teknium 9b05f0df55 refactor(doctor): split config and host-platform checks into doctor_config.py / doctor_platform.py 2026-09-02 15:56:21 -07:00
Teknium 53a24170ab refactor(auth): split Nous Portal OAuth flow into hermes_cli/auth_nous.py 2026-09-02 15:46:40 -07:00
Teknium 3c7cd69abb refactor(tools_config): extract cua-driver installer cluster into tools_config_cua 2026-09-02 15:43:44 -07:00
Teknium 5608fe7cdf refactor(web_server): extract chat/PTY/WS-auth cluster to web_server_chat 2026-09-02 15:36:53 -07:00
Teknium 88b74d6ef0 Merge branch 'simp/hclib' into simp/integration 2026-09-02 15:05:48 -07:00
Teknium 7df99a7787 test(hclib): update tests for moved/removed symbols 2026-09-02 15:03:42 -07:00
Teknium 24b9d18448 Merge branch 'simp/cli-models' into simp/integration
Conflict in hermes_cli/model_setup_flows.py: took simp/cli-models's
simplified version and re-applied simp/cli-main's 2-line change (import
_current_reasoning_effort/_set_reasoning_effort from hermes_cli.setup
instead of hermes_cli.main, since cli-main removed the main.py copies).
simp/cli-models had dropped those two helpers from hermes_cli.setup.py as
dead (their only base caller was main.py's own copy); restored them there
so the cli-main import path resolves.
2026-09-02 15:01:11 -07:00
Teknium 1593bee0c3 Merge branch 'simp/cron' into simp/integration 2026-09-02 14:19:40 -07:00
Teknium 477dc3de55 Merge branch 'simp/gw' into simp/integration 2026-09-02 14:19:40 -07:00
Teknium 583e49ab4f Merge branch 'simp/hclia-transports' into simp/integration 2026-09-02 14:19:38 -07:00
Teknium 81ddc99705 Merge branch 'simp/cli-update' into simp/integration 2026-09-02 14:19:37 -07:00
Teknium 50be642e52 Merge branch 'simp/cli-misc' into simp/integration 2026-09-02 14:19:37 -07:00
Teknium cc9ef82355 refactor(gateway-cli): dispatch-table subcommands and shared service helpers in hermes_cli/gateway.py
hermes_cli/gateway.py (9178 -> 6906):
- `hermes gateway` subcommand routing: if/elif chain -> _GATEWAY_SUBCOMMANDS dispatch table of
  _cmd_* handlers; _stop_installed_service and _refuse_from_inside_gateway unify the
  stop/restart/uninstall service-stop and self-target guards.
- One systemd unit template; _systemctl_show, systemctl reset+action, is-active probes,
  _installed_service_kind ladder (3 sites), launchctl-list PID probe, launchd bootstrap+kickstart
  helper, legacy-unit removal loop, loop-tick witness ping, ps/wmic line parsers, planned-stop
  marker helper, _CAPTURE_TEXT subprocess kwargs (21 sites), _gw_windows() accessor (15 lazy
  imports).
- run_gateway startup helpers, reaper exclusion set, Windows process listing extracted; wizard
  service actions and platform-setup prompts unified (_setup_service_action).
- Dead: _windows_scheduled_task_running (test-only), _container_systemd_operational (inlined),
  whatsapp/email/matrix built-in status branches (those platforms are registry plugins).
- try/except-pass around single statements -> contextlib.suppress; comments/docstrings compacted
  to the WHY (bounded_probe_run vs subprocess.run on Windows, pythonw launcher-stub filtering,
  raw-record vs validated-probe exclusion in the orphan reaper, Scheduled-Task Ready-vs-Running).
- `hermes gateway --help` byte-identical.
2026-09-02 13:31:44 -07:00
Teknium 75bafc197a refactor(plugins): unify provider registration and unload paths in hermes_cli/plugins.py
hermes_cli/plugins.py (7193 -> 4961):
- _register_scoped_provider: one body for the 8 scope-keyed register_*_provider methods;
  _track_callback/_track_mapping_entry unify manager-mapping lease tracking (4 sites);
  _track_scoped_registration for the ownership ledger. Public method names, signatures,
  return values and warning strings unchanged.
- Manifest v2 type checks table-driven (_manifest_field_of_type); _unload_scoped split into
  _unload_target_keys + _reset_after_unload_all; _gate_manifest/_record_placeholder out of
  _discover_and_load_inner; _track_tool_override_policy/_attribute_registrations out of
  _load_plugin_scoped; bounded hook worker lifted into _run_hook_callback_bounded;
  prompt-section rendering extracted; resolve_pre_tool_block delegates to
  _dispatch_pre_tool_call_hooks; _evict_modules (3 sites); _remove_name_if_unowned and
  _nowait_plugin_set unify unowned-name cleanup and *_nowait probes.
- Dead (zero references outside their own test): unload_plugins, has_portable_mcp_servers,
  _classify_entrypoint_kind, _reset_event_bus, get_plugin_subscriptions,
  get_telegram_handler_factories, pass-through _restore_* helpers; _env_enabled is now an alias
  of utils.env_var_enabled (plugins/memory still imports it).
- Docstrings/comments compacted; contract sentences and rationale kept (multi-profile ledger
  keying, persistent auth-provider registration, capability declaration is not a grant).
2026-09-02 13:31:44 -07:00
Teknium 401efb0b2b refactor(cli): decompose _cmd_update_impl into phase helpers, unify duplicated update helpers, compact comments
hermes_cli/update_cmd.py 11217 -> 10204 LOC; _cmd_update_impl 2797 -> 521 LOC.

Decomposition (behavior-neutral, AST free-name verified) — _cmd_update_impl is
now a thin orchestrator calling, in order:
  _clear_windows_venv_holders_or_exit -> _prepare_checkout_for_update
  (-> _CheckoutPlan) -> _repair_current_checkout | _pull_updates ->
  _sync_python_dependencies_after_pull -> _run_post_update_maintenance ->
  _restart_gateway_fleet_after_update (-> _GatewayRestartOutcome;
  _restart_systemd_gateway_units) -> _resume_windows_gateways_and_merge_outcome
  -> _verify_fleet_after_update. _resolve_manage_cmd hoisted to module level.

Unified helpers: _git_run (49 captured git subprocess.run sites),
_systemctl / _systemctl_reset_and_restart (17 sites + 2 pairs),
_sweep_bytecode_after_update (3x), _print_bundled_skills_sync_report (ZIP+git),
_ensure_venv_pip (ZIP+git), _self_and_non_gateway_ancestor_pids (2x),
_record_update_step (4x), _write_gateway_update_exit_code for the 3 inline
".update_exit_code" writes. Dropped duplicate nested copies of
_wait_for_service_active/_service_restart_sec/_print_items, dead
upstream_exists, a dead if/pass branch and unused imports.

Comments/docstrings hand-compacted (236 blocks, AST-identical with docstrings
normalized); rationale/invariant sentences kept.

Tests: source-inspection guards repointed to the helper that now owns the code
(test_update_self_lock, test_update_fleet_check_fail_closed,
test_update_apply_shallow_count); new regression test drives the real Windows
resume/merge helper.
2026-09-02 13:30:54 -07:00
Teknium 3c7069bdcb refactor(gateway): dead-code removal, helper unification, defensive-layer collapse and rationale-preserving comment compaction across 40 modules
authz_mixin, browser_control_broker, delivery, delivery_ledger, display_config, drain_control,
hosted_room_links/peer/policy_checkpoint, hosted_rooms, platform_registry, relay/__init__,
relay/ws_transport, run.py and slash_commands.py (comments), session_context, session_state,
streaming_tts_consumer, turn_lease and small modules.

- HostedRoomPolicyCheckpoint._apply_event -> per-kind handler table
- WebSocketRelayTransport._handle_frame -> frame-handler table
- GatewayAuthorizationMixin: unified adapter setting/flag/extra readers
- dead symbols removed (verified zero references): RoomLinkProbe/select_room_link,
  relay_bot_username, is_restart_loop_tripped, debug_rows, DeadTargetRegistry.all_dead,
  BrowserControlBroker.detach_owner/_prune_tickets, StreamingTTSConsumer.started/_enqueue_done/
  _iter_stream_chunks/_next_stream_chunk, RecoverableHandleCache.status_for, _auth_env,
  _copy_default_catalog, _parse_timestamp_prefix, _present_* helpers, _send_result_error_kind,
  _truthy_env, SessionFieldView/TurnLeaseTokenView dunder shims, and their orphaned tests.
- lost WHY/invariant text from the earlier compaction restored compactly (541 hunks audited)
2026-09-02 13:30:50 -07:00
Teknium 00cd12f7e4 refactor(agent/relay): fold relay_llm attempt scaffolding into _ManagedAttempt; dedupe runtime session/scope paths; drop dead relay_runtime module-level shims 2026-09-02 13:29:48 -07:00
Teknium 9dab23c5a6 refactor(cli): simplify hermes_cli/commands.py registry, gateway collectors and completer
- Drop dead code: _telegram_effective_priority, _prioritize_telegram_menu_commands,
  discord_skill_commands, _TG_NAME_LIMIT/_clamp_telegram_names compat aliases,
  the empty _SLACK_PRIORITY_ALIASES pinning pass (and tests that only pinned them).
- Unify the Telegram and Discord skill collectors behind _iter_gateway_skills
  (one eligibility/root-matching implementation) and _truncate_desc.
- Table-drive Telegram menu priority modes (_TELEGRAM_PRIORITY_TIERS) and the
  completer's per-command dynamic completions (_DYNAMIC_COMPLETIONS).
- Unify path and @file:/@folder: directory-listing completions (_dir_completions),
  command-completion construction (_short_desc), /tools candidate rows, the
  Slack canonical/alias passes and the derived COMMANDS/COMMANDS_BY_CATEGORY loops.
- Compact docstrings/comments, keeping every rule, invariant and rationale.
Behavior-neutral: registry-derived outputs, menus, manifests and completions
byte-identical against origin/main on a fixture sweep.
2026-09-02 13:29:45 -07:00
Teknium b81ab901f3 refactor(cli): extract the last 16 inline parser groups from main() into hermes_cli/subcommands/
moa, fallback, worktree, browser, secrets, egress, migrate, whatsapp-cloud,
checkpoints, bundles, curator, pets, journey, computer-use, sessions and
completion each become a build_<group>_parser() builder. Closure handlers
that only closed over their own parser moved verbatim; sessions/completion
take the handler by injection. --help/usage/defaults byte-identical for all
399 parsers in the tree (in-process dump before/after).
2026-09-02 13:29:44 -07:00
Teknium 5212a3077d refactor(cli): dedupe model_setup_flows boilerplate
model_setup_flows.py (3313 -> 2848):
- _load_config_model_section, _begin/_commit_model_config, _ensure_flow_api_key,
  _pick_model_or_prompt, _run_login, _models_dev_merged, _copilot_model_list,
  _show_curated replace ~15 copies of config-save / api-key / picker boilerplate.
- _gemini_tier_ok and _api_key_provider_model_list lift the two inline blocks
  out of _model_flow_api_key_provider; five-way provider branch -> early returns.
- Comments compacted, keeping every rationale (Bedrock geo routing, key_env
  hygiene, discover_models semantics, Nous free/paid partition, etc.).

Also drops two tests that only asserted the existence of setup.py helpers
removed in the next commit.
2026-09-02 13:29:29 -07:00
kshitijk4poor a3d33fe22f fix(copilot): single-flight the token exchange and close abandoned responses
Follow-up to the off-loop move: once the credential-pool handlers run on
worker threads, the dashboard's periodic /api/credentials/pool polls can
overlap, and during a DNS outage each poll would have started its own
exchange and abandoned its own hung resolver thread.

- Per-fingerprint threading.Lock around the exchange: concurrent callers
  wait on the one in-flight attempt, then hit the positive or negative
  cache (bounded worker count, no duplicate network calls).
- _urlopen_bounded: when the hard cap fires and the abandoned worker later
  succeeds, close the HTTPResponse instead of leaking the socket.
- Tests (none shipped with the original PR): hard cap + late-close,
  single-flight success and failure paths, and the pool endpoint running
  off-loop / keeping the loop responsive under a 200 ms blocking read.
2026-09-03 01:57:09 +05:30
kshitijk4poor 2b4e70ec07 refactor(cli): use the router's run_in_threadpool alias; offload profile model write
The module already binds run_in_threadpool (used by list_profiles_endpoint)
and every sibling router uses the same starlette helper; the nine new
loop.run_in_executor(None, _run) sites now go through that alias so the
file has one offload idiom. Behaviour-identical (both hand the callable to
a worker thread).

Also sweeps the one endpoint the PR left synchronous:
update_profile_model_endpoint's _write_profile_model reads and rewrites
the profile's config.yaml on the event loop.
2026-09-03 01:56:47 +05:30
briandevans beb2e91d04 test(cli): cover the profiles router off-loop sweep
Two assertions per offloaded site:

- a loop probe, where the stubbed callee records whether an event loop is
  running in its own thread — the idiom already used by
  tests/hermes_cli/test_cron_dashboard_off_loop.py; and
- a concurrency proof, where the stubbed callee blocks on a threading.Event
  while an unrelated request is timed. On the unfixed handlers that request
  waits out the whole block; served off the loop it returns in
  milliseconds.

The concurrency proof needs a single event loop across requests, so the
client fixture enters the TestClient context manager: that pins one
blocking portal for the whole fixture, where a bare TestClient(app) would
spin up a fresh loop per request and pass even unfixed.

Also covers the status-code mapping through the executor hop (404 on a
missing profile, 400 on a rename collision, 404 from the resolve that stays
on the loop ahead of describe-auto) and the _MISSING sentinel cases: a
desktop.json holding `null` still reports exists=true, an absent one
reports exists=false, and an empty SOUL.md is still distinguishable from a
missing one.

The client fixtures read web_server._SESSION_TOKEN from the module rather
than pinning a literal. web_server resolves that token once at import, so
whichever test file imports it first fixes the value for the session and a
later monkeypatch.setenv is silently ignored — two files hardcoding
different tokens would 401 depending on collection order.
2026-09-03 01:56:47 +05:30
briandevans f786c8699b test(proxy): cover the retry credential blocking the proxy event loop
Extends the off-loop suite to the third and last blocking method on the
`UpstreamAdapter` contract, the 401/429 rotation.

As with the two existing pairs, the primary assertion is **thread identity**,
not latency: a latency assertion measured by an HTTP client on the blocked
loop is vacuous, because the client's own timer cannot advance until the
block ends and it therefore reports a fast response on provably frozen code.

  * `test_get_retry_credential_runs_off_the_event_loop` records
    `threading.get_ident()` inside the fake adapter and compares it to the
    loop thread, and checks the rotation still works end to end (rejected
    bearer forwarded first, rotated bearer second).
  * `test_event_loop_keeps_running_while_the_retry_credential_resolves`
    samples a loop-side heartbeat counter from inside the stalled adapter. On
    the unfixed handler it records exactly 0 loop iterations across a 0.5s
    rotation.
  * `test_retry_credential_failure_still_returns_the_upstream_rejection`
    guards the error contract the change must leave alone: a raising rotation
    is still swallowed and the upstream's own 401 is streamed back, with no
    second forward.

A new `_build_rejecting_upstream` harness drives the `status in {401, 429}`
branch by rejecting every bearer except the rotated one.
2026-09-03 01:56:32 +05:30
briandevans 4b2f16d828 test(proxy): cover health responsiveness while the auth store is locked
Extends `test_proxy_off_loop.py` with the `/health` half, using the same
two-assertion shape as the credential tests:

- `test_is_authenticated_runs_off_the_event_loop` compares the thread the
  adapter's `is_authenticated` ran on against the loop thread. Before the
  fix they are the same ident.
- `test_event_loop_keeps_running_while_health_resolves_auth_state` reads a
  loop-side heartbeat counter sampled by the adapter across its own stall.
  Before the fix exactly 0 iterations run across 0.5s.

Both also assert the response is unchanged (`200`, `authenticated: true`),
so the offload cannot quietly alter what `/health` reports.
2026-09-03 01:56:32 +05:30
briandevans d4c2c74ece test(proxy): cover credential resolution blocking the proxy event loop
Adds `tests/hermes_cli/test_proxy_off_loop.py`, mirroring the harness in
`test_proxy.py`: the proxy and a fake upstream run as real aiohttp
servers on ephemeral ports under a single `asyncio.run`, guarded by
`pytest.importorskip("aiohttp")` — no pytest-aiohttp dependency.

The primary assertion is thread identity, not latency. A latency
assertion measured with an HTTP client on the blocked loop is vacuous:
the client's own timer cannot advance until the block ends, so it reports
a fast response on code that was provably frozen.

- `test_get_credential_runs_off_the_event_loop` records
  `threading.get_ident()` inside the adapter and compares it to the loop
  thread. Before the fix both are the same ident.
- `test_event_loop_keeps_running_while_credentials_resolve` runs a
  heartbeat task on the loop and has the adapter sample its counter on
  entry and exit, so the reading is taken from the loop rather than
  through a client that shares it. Before the fix exactly 0 iterations
  run across a 0.5s stall; after it, ~50.
- `test_credential_failure_still_maps_to_401` pins the error contract
  across the change of call form. It is deliberately not in the
  red-before set — it guards behaviour the fix must leave alone.
2026-09-03 01:56:32 +05:30
Teknium 10088c569e fix(update): hermes update no longer hangs on a GitHub username prompt
GitHub answers anonymous fetches with HTTP 401 during outages (and for
renamed/private repos). git then prompts `Username for 'https://github.com':`
on the inherited terminal and `hermes update` sits there — users read it as
Hermes demanding a GitHub login.

Every network git call in the updater (fetch/pull/push, apply + --check +
fork sync) now runs with GIT_TERMINAL_PROMPT=0 / stdin=DEVNULL, so the 401
fails fast into the fetch-failure classifier, which now reports it as a
GitHub-side rejection (likely outage) rather than blaming the user's
credentials. Credential helpers/askpass are left configured so private-fork
origins still authenticate.

Live repro: PTY-attached update --check against a 401 origin hung 15s+ on
the prompt before; exits rc=1 in 0.2s with the diagnosis after.

Same class as #73751 (@Frowtek, pre-main.py decomposition); passive banner
half salvaged from #101421 (@RobbertC5).
2026-09-02 12:34:26 -07:00
RobbertC5 6e775907d7 fix(cli): keep passive update checks noninteractive 2026-09-02 12:34:26 -07:00
Teknium 413b6ba3dd Port from google-gemini/gemini-cli#28792: harden internal git env
(cherry picked from commit 09bb9c3b2f20f22bdb1688b6efcb78b24f1f19d6)
2026-09-02 10:33:43 -07:00
Teknium 457d9b73f6 test/docs(nous): trim keepalive tests to the two invariants; document nous.keepalive_interval_seconds 2026-09-02 10:22:30 -07:00
olopez25 c6c8c74c70 Move the keepalive interval to config.yaml and tighten the schedule guard
Addresses review feedback on #84928.

The tick interval was exposed as HERMES_NOUS_KEEPALIVE_INTERVAL_SECONDS.
AGENTS.md reserves .env for credentials and puts behavioural thresholds in
config.yaml, so the knob moves to `nous.keepalive_interval_seconds`,
following the existing `vertex:` section's precedent for non-secret
provider settings. The env var is dropped rather than bridged: it was never
released, so nothing depends on it. Adding a key to a new section is handled
by the deep-merge, so no _config_version bump is required.

test_keepalive_interval_fits_inside_the_token_lifetime asserted
`900 < 899 * 4 - 120`, which is true for any realistic interval and could
never fail. It also tested the wrong value: the configured constant is only
a ceiling, while the schedule that ships is the derived tick. Replaced with
an assertion over the derived tick for each observed lifetime, which does
fail if the derivation constants regress -- verified against both
TICKS_PER_LIFETIME=1 and MIN_INTERVAL_SECONDS=5000.

Also adds coverage for an unreadable config.yaml, which must fall back to
the module default rather than take the keepalive thread down.

pytest tests/hermes_cli/test_nous_auth_keepalive.py -> 9 passed

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-02 10:22:30 -07:00
olopez25 fd05b97913 Derive Nous keepalive tick from the issued credential lifetime
Follow-up to the interval fix: ticking faster narrowed the gap but did not
close it, and the hardcoded interval was wrong for short-lifetime accounts.

Two independent problems, both needed:

1. Lifetime is not a constant. Installs have been observed issuing ~3594s
   and ~899s (see #35752, which reports expires_in: 899 while this account
   reports 3594). Any hardcoded interval is right for one and wrong for the
   other. The tick now derives from the lifetime the server actually issued,
   capped by the configured interval and floored at 60s. The access token and
   the invoke agent key carry separate lifetimes, so the shorter one governs.

2. Ticking faster alone never closes the gap. The refresh only fires once a
   credential is within ACCESS_TOKEN_REFRESH_SKEW_SECONDS (120s) of expiry,
   so a tick spaced wider than that window steps straight over it. With a 900s
   tick against a 3594s lifetime the last tick before expiry still saw 894s
   remaining, declined to refresh, and the credential died 6s before the next
   one. The keepalive now asks "will this outlive my next tick?" (tick + skew)
   rather than reusing the request path's bare skew.

Measured against both observed lifetimes:

  lifetime 3594s: was never refreshed proactively; now refreshes 900s early
  lifetime  899s: was never refreshed proactively; now refreshes 227s early

The lifetime is re-read every pass rather than cached, since it can change
when the account, plan, or server-side policy does -- exactly the case the
keepalive exists to cover.

min_access_ttl_seconds is threaded through as an optional parameter, so the
request path keeps its existing 120s behaviour and only the keepalive widens
the window.
2026-09-02 10:22:30 -07:00
olopez25 a301d19ba4 Fix reactive Nous 401s by ticking keepalive inside the token lifetime
Nous Portal access tokens carry a one-hour lifetime, and the keepalive only
refreshes once a token is within ACCESS_TOKEN_REFRESH_SKEW_SECONDS (120s) of
expiry. The tick interval was 6 hours, so it could only land inside that
2-minute window by coincidence. In practice every hour rolled over untouched
and the next inference call paid a 401 plus a re-auth round trip.

Observed on a production install: 71 "refreshed Nous runtime credentials
after 401, retrying" events across current logs.

Drop the default interval to 15 minutes, which gives four ticks per token
lifetime and leaves ample margin under the TTL-minus-skew ceiling of 3480s.
Add HERMES_NOUS_KEEPALIVE_INTERVAL_SECONDS so the interval is tunable without
a source edit, matching the existing HERMES_NOUS_TIMEOUT_SECONDS convention.
Zero still disables the thread.

Callers pass no interval, so the default is what actually shipped; the
signature now resolves at call time rather than binding the constant at
import.
2026-09-02 10:22:30 -07:00
Teknium 0cbc6e37ac test/docs: trim seam tests to invariants, document create_client and external-process fields
Cuts the 41 contributor tests down to 8 pinning the before/after contracts
(out-of-tree provider resolves end to end, copilot-acp unchanged, broken
plugin falls through, flat-install discovery + non-provider kinds untouched).
Adds the create_client hook and process_* fields to the model-provider
plugin developer guide.
2026-09-02 09:57:39 -07:00
Alexander Prendota 1bd0b8ed41 feat(providers): let an external-process provider ship out of tree
An external-process provider is an agent CLI Hermes drives over stdio rather
than an HTTP endpoint. Three things about it were spelled out for one vendor,
and each was a hard stop for any other:

* ``resolve_provider()`` gates on ``PROVIDER_REGISTRY``. Its auto-extend from
  ``providers/`` covered api-key providers only, so an external-process profile
  never entered it and ``hermes -m <that provider>`` died with "Unknown
  provider" before a client was ever built.
* ``resolve_runtime_provider()`` keyed the external-process branch on the
  literal ``"copilot-acp"``, so anything else silently fell through to the
  OpenRouter default instead of its own runtime.
* ``resolve_external_process_provider_credentials()`` hardcoded the binary
  (``copilot``), the argv (``--acp --stdio``), the env var names and the
  placeholder api_key — so a third-party provider would have been handed
  another vendor's CLI.

Now the profile carries what only the provider knows — ``process_command``,
``process_args``, ``process_command_env_vars``, ``process_args_env_var`` — and
the three core paths key on ``auth_type == "external_process"`` instead of a
name. copilot-acp's values move into its profile verbatim, so
``HERMES_COPILOT_ACP_COMMAND`` / ``COPILOT_CLI_PATH`` /
``HERMES_COPILOT_ACP_ARGS`` and its ``copilot-acp`` api_key placeholder behave
exactly as before; the new tests assert that alongside the out-of-tree case at
every step.

The error for a missing binary now names the provider and its own env override
instead of telling every user to install GitHub Copilot CLI.

Co-Authored-By: Junie <junie@jetbrains.com>
2026-09-02 09:57:39 -07:00
unsupportedpastels 5487222658 fix(models): live Copilot catalog for CLI-login users; unbreak picker-flag-empty catalogs
Three gaps between the copilot-acp picker row and what the user's
subscription actually serves (reported: picker showed the stale curated
list while the Copilot CLI offered Sonnet 5 / Opus 5 / GPT-5.6):

1. _resolve_copilot_catalog_api_key() never looked at the Copilot CLI's
   own token store (~/.copilot/config.json copilotTokens). A user whose
   only credential is 'copilot login' got no catalog key, the live fetch
   401'd, and copilot-acp silently fell back to the stale curated list.
   Add it as resolution source 3, JSONC-tolerant, with each candidate
   validated and exchanged like pool entries.

2. The existing credential-pool branch unpacked exchange_copilot_token()
   into two names, but it returns (api_token, expires_at, base_url) —
   the ValueError was swallowed by the enclosing except, disabling that
   entire resolution path. Latent since the base_url return was added.

3. GitHub now returns model_picker_enabled: false for EVERY model on
   some accounts/token types, so honoring the flag rejected the whole
   live catalog. Treat the flag as a display hint: when it empties the
   result, refilter without it (chat/endpoint checks still exclude
   embeddings and non-chat rows).

Verified live: catalog resolves 44 models for a copilot-login-only
account, matching the CLI's own picker (claude-sonnet-5, claude-opus-5,
gpt-5.6-sol/terra, gemini, kimi).
2026-09-02 20:51:07 +05:30
unsupportedpastels 6b2d32d3f6 fix(picker): keep signed-in copilot-acp visible in explicit-only desktop pickers
Two follow-up gaps found by actually running 'copilot login' end-to-end:

1. The CLI (without an OS keychain) stores its token in
   ~/.copilot/config.json under copilotTokens — a JSONC file with
   //-comment header lines. Add it as an auth-evidence source in
   _external_process_auth_evidence(), parsed comment-tolerantly and
   counting only a non-empty copilotTokens map (config.json exists after
   first launch even when logged out).

2. The desktop chat picker requests explicit_only rows, and
   _filter_explicit_provider_rows() dropped copilot-acp because a CLI
   login leaves no trace in active_provider, model.provider, or env vars
   — exactly the Anthropic-OAuth carve-out case. Keep external_process
   rows when their CLI credentials are verified (auth_verified), while
   still dropping ambient executable-on-PATH-only rows so the filter's
   narrower contract holds.

Net effect: after 'copilot login', copilot-acp appears in the desktop
picker and the Accounts card reads signed in; a machine with only the
binary installed keeps today's hidden-until-configured behavior.
2026-09-02 20:51:07 +05:30