+ jHj=Rt^RIHt^RIt^RIt^RIt^RIt^RIHtH t ^RI H t ^RI H t Ht^RIHtHtHtRtRtRt] !R R 44tR R lt!R R4t]P2R4tRRltR#)aEmbedded kanban dispatcher: settings resolution and per-tick board work. ``GatewayKanbanWatchersMixin._kanban_dispatcher_watcher`` owns the loop, the singleton lock and the health telemetry; everything that only needs the ``kanban_db`` module and the resolved settings lives here. ) annotationsN)asdict dataclass)Path)AnyOptional) _board_slugs_positive_int_settingloggerc^RIHpV#)kanban_db_connect) hermes_clirr s G/home/ubuntu/.hermes/hermes-agent/gateway/kanban_watchers_dispatcher.py_kbcrs , c^RIHpV#)r kanban_db_dispatch)rrrs r_kbdrs - rcl]tRt^!t$RtR]R&R]R&R]R&R]R &R]R &R ]R &R ]R&R]R&RtR#)_DispatcherSettingszE``kanban.*`` dispatch settings, read once at boot (restart to apply).floatintervalr max_spawnz Optional[int]max_in_progressint failure_limitstale_timeout_secondsboolreconcile_orphansz Optional[str]default_assigneemax_in_progress_per_profileN)__name__ __module__ __qualname____firstlineno____doc____annotations____static_attributes__r$rrrr!s3OON""##!..rrc$V^8dQhRRRRRR/#) kanban_cfgdictkbrreturnrr$)formats"r __annotate__r3/s'GGTGsG?RGrc \VPR^<4;'g^<4p\ VR4pVPR4pVe\P!RV4\VR4p\4PV4pVfVe\P!R V4VPR VP4p\V4pV^8d.\P !R WaP4VPpVPR ^4p\T;'g^4p VPR4;'gRP4;'gRp V 'd\P!RV 4\VVVVV \VPRR44V \VR4R7# \\3d-\P !RTPR44RpELi;i \\3d2\P !R YaP4TPpELei;i \\3d\P !RT4^p EL:i;i)zAParse and log the dispatcher settings in their established order.dispatch_interval_secondszIkanban dispatcher: invalid dispatch_interval_seconds=%r, using default 60gN@g?rNzkanban dispatcher: max_spawn=%srzkanban dispatcher: kanban.max_in_progress unset; using memory-derived default max_in_progress=%d (set kanban.max_in_progress in config.yaml to override)rzDkanban dispatcher: invalid kanban.failure_limit=%r; using default %dzGkanban dispatcher: kanban.failure_limit=%r is below 1; using default %ddispatch_stale_timeout_secondsz^kanban dispatcher: invalid kanban.dispatch_stale_timeout_seconds=%r; disabling stale detectionr"zZkanban dispatcher: default_assignee=%r (unassigned ready tasks will route to this profile)r!Tr#)rrrrrr!r"r#)rget ValueError TypeErrorr warningmaxinfor rresolve_max_in_progressDEFAULT_FAILURE_LIMITrstriprr ) r.r0rrreffective_max_in_progressraw_failure_limitr raw_stalerr"s && r_resolve_dispatcher_settingsrD/sC(CRHNNBO 8S!H{+I 5yA ,J8IJO $ > > O#<#H  F &  #8P8PQ1-. q`(*B*B D00 ?CI" #INN 3#'9:@@bGGIQQT 23C E 1#3z~~.A4HI)%:*Fc$d  m  "b!~~&AB D4 z "1](*B*B D00 1 z ""34= ? !"s@F/F/ G/ H4 H4/9G,+G,/>H10H14*I"!I"c]tRt^ytRtRtRRltRRltRRltR R lt R R lt R Rlt RRlt RRlt RRlt]RRl4tRtR#)_KanbanDispatchera[Per-tick board work for the embedded dispatcher (runs in worker threads). Boards are enumerated every tick so a board created mid-run is picked up without a restart. Corrupt-looking board DBs are quarantined per fingerprint and retried after ``CORRUPT_BOARD_RETRY_AFTER_SECONDS``: transient WAL/open races can look like "malformed" for one tick. i,c$V^8dQhRRRRRR/#)r-r0rsettingsrr1Noner$)r2s"rr3_KanbanDispatcher.__annotate__s'gg3g*=g$grc ,WnW n/VnR#N)r0rHdisabled_corrupt_boards)selfr0rHs&&&r__init___KanbanDispatcher.__init__s df$rcV^8dQhRR/#)r-r1listr$)r2s"rr3rJs%%d%rc ,\VP4#rL)rr0)rNs&rr_KanbanDispatcher._board_slugssDGG$$rc V^8dQhRRRR/#)r-slugstrr1z"tuple[str, int | None, int | None]r$)r2s"rr3rJs : : :1S :rc ^RIHpVP4;_uu_4VPP V4pRRR4\ XP 4P44pVP4pYEPTP3# +'giLa;i \d\ X4pLTi;i \d TRR3u#i;i)r  kanban_dbN) rrZpin_first_board_resolutionr0kanban_db_pathrW expanduserresolve ExceptionstatOSError st_mtime_nsst_size)rNrV_kbpathresolvedr`s&& rboard_db_fingerprint&_KanbanDispatcher.board_db_fingerprints/  + + - -77))$/D. !4??,4467H *99;D**DLL99. - !4yH ! *dD) ) *s/B'B',C B$ 'C?CCCc V^8dQhRRRR/#)r-excr_r1r r$)r2s"rr3rJsrrYr4rrc a\S\4P4'dR#\S\P4;'dC\ ;QJd#V3Rl\ 4F 'gK R# R#!V3Rl\ 44#)Tc3Z<"TF q\S4P49xK" R#5irL)rWlower).0mrjs& r >_KanbanDispatcher.is_corrupt_board_db_error..s"=q]pXY3s8>>CS>S]ps(+F) isinstancerKanbanDbCorruptErrorsqlite3 DatabaseErrorany_CORRUPT_DB_MARKERS)rNrjs&fris_corrupt_board_db_error+_KanbanDispatcher.is_corrupt_board_db_errors[ c4666 7 7#w445qq##=q]p=q##q#q#=q]p=q:qqrc$V^8dQhRRRRRR/#)r-rVrW fingerprinttupler1r r$)r2s"rr3rJs!s4rc ZVPPV4pVfR#VwrE\P!4V, pWB8XdW`P8dR#WB8Xd\ P !RW4M\ P !RV4VPPVR4R#)zFReturn False while *slug* stays quarantined; lift (and log) otherwise.NTFzdkanban dispatcher: board %s database fingerprint unchanged after %.0fs quarantine; retrying dispatchz?kanban dispatcher: board %s database changed; retrying dispatch)rMr8time monotonic!CORRUPT_BOARD_RETRY_AFTER_SECONDSr r=pop)rNrVr{disabled_entrydisabled_fingerprint disabled_atages&&& r_quarantine_lifted$_KanbanDispatcher._quarantine_lifteds5599$?  !,:)nn,  .39_9_3_  . KKDEI P KKY[_ ` $$((t4rc V^8dQhRRRR/#)r-rVrWr1zOptional[object]r$)r2s"rr3rJs&!&!&!0@&!rc  HRpVPV4pVPW4'gR#\VP4P 4UUu/uFwrEVR8wgKWEbK ppp^RIHpVP4;_uu_4\4PVR7p\4P!V3RV/VBuuRRR4Ve=\P!\4;_uu_4VP4RRR4##uuppi +'gi#;i +'giEM=;i \Ed,pTP!T4'dT\"P$!43TP&T&\(P*!RY^,4Rp?TeR\P!\4;_uu_4TP4RRR4R# +'giR#;iR#\(P,!RT4Rp?TeR\P!\4;_uu_4TP4RRR4R# +'giR#;iR#Rp?ii;iTeR\P!\4;_uu_4TP4RRR4R# +'giR#;iR# TeP\P!\4;_uu_4TP4RRR4i +'gii;ii;i)zRun one dispatch_once for a specific board. The per-board DB is opened explicitly so boards never share a connection or claim across each other. NrrYboardra kanban dispatcher: board %s database %s is not a valid SQLite database; pausing dispatch for this board until the file changes, the gateway restarts, or the quarantine timer expires. Move or restore the file, then run `hermes kanban init` if you need a fresh board.z*kanban dispatcher: tick failed on board %s)rgrrrHitemsrrZr[rconnectr dispatch_once contextlibsuppressr_closerxr~rrMr error exception) rNrVconnr{kvkwargsrdrjs && rtick_once_for_board%_KanbanDispatcher.tick_once_for_boardsF //5 &&t99#)$--#8#>#>#@T#@41AO$!$#@T ! 4//11v~~D~1v++DGGG21$((33JJL43 5U643'211 --c226A4>>CS5T,,T2 O a. ((33JJL4333    I4 P((33JJL4333  2$((33JJL4333 t((33JJL4333 s D #D -D: 7D& D:2D D# & D7 1D:5K 7D:: I1AI,K G$$ G5 :I,K K ;I I' ,I11K J66 K 'L!3L  L! L  L!cV^8dQhRR/#)r-r1z"list[tuple[str, Optional[object]]]r$)r2s"rr3rJsXX=Xrc lVP4Uu.uFqVPV43NK up#uupi)z>Run one dispatch_once per board. Returns (slug, result) pairs.)rr)rNrVs& r tick_once_KanbanDispatcher.tick_onces3CGCTCTCVWCV4t//56CVWWWs1cV^8dQhRR/#)r-r1r r$)r2s"rr3rJsrc  p\4pVP4p^RIHpVP 4;_uu_4VP 4FpRp\ 4PVR7pVPV4'g V'dcVPV4'dLVe<\P!\4;_uu_4VP4RRR4RRR4R#Te>\P!\4;_uu_4TP4RRR4KK RRR4R# +'giLj;i \d]TfEK\P!\4;_uu_4TP4RRR4EKT +'giEKf;ii;i +'giEK;i TeP\P!\4;_uu_4TP4RRR4i +'gii;ii;i +'giR#;i)uIs there a ready+assigned+unclaimed task on ANY board the dispatcher would spawn for? Control-plane lanes (e.g. ``orion-cc``) are pulled by terminals via ``claim_task`` and never spawnable — a queue full of those is "correctly idle", not "stuck". The review column is probed only when review dispatch is on (same gate as the dispatcher): a task waiting for a human reviewer is idle, not stuck. rYNrTF)rreview_dispatch_enabledrrZr[rrrhas_spawnable_readyhas_spawnable_reviewrrr_r)rNkbd _review_proberdrVrs& rready_nonempty _KanbanDispatcher.ready_nonemptysf335 /  + + - -))+ )6>>>5D..t443KcKcdhKiKi#''00;; JJL<. -''00;; JJL<;(,.<; !''00;; JJL<;;; <;;''00;; JJL<;;;(. -sH$8E  E !&H$D:  H$+&H$F7 "H$: E H$ F4 G H$ "H$F  H$ F0) H$3F4 4G 7 G  H$ 'H! 3H  H! H H! !H$$ H5 c V^8dQhRRRR/#)r-auto_decompose_per_tickrr1r$)r2s"rr3rJs&&3&3&rc  ^RIHp^p^p^RIHp\ 4;_uu_4TP4;_uu_4TP4FpYA8dM\PPR4pT\PR&TP4p T F,p YA8dM%T^, pYPPY'T 4, pK. Tf#\PPRR4KT\PR&K RRR4RRR4T# \d#p\P!RT4Rp?^#Rp?ii;i \d$p\P!RYs4.p Rp?LRp?ii;i Tf"\PPRR4iT\PR&i;i +'giL;i +'giT#;i)zAuto-decompose up to N triage tasks across all boards into ready workgraphs. Runs before dispatch fans out; the per-tick cap keeps a bulk triage load from burst-spending the aux LLM. Returns the number decomposed. )kanban_decomposez3kanban auto-decompose: import failed (%s); skippingNrYHERMES_KANBAN_BOARDz>kanban auto-decompose: list_triage_ids failed on board %s (%s))rrr_r r;rZ_default_profile_secret_scoper[rosenvironr8list_triage_idsdebug_decompose_oner) rNr_decomprj attempted successesrdrVprev_env triage_idstids && rauto_decompose_tick%_KanbanDispatcher.auto_decompose_ticks   >  / * , ,c.L.L.N.N))+7::>>*?@E8  *$?!!Q !%8%8%LL *  ' '>>g/// KKH$UXY`YjYjUk l KKTVZ `    RTW X sB55"CC)rMr0rHN)r%r&r'r(r)rrOrrgrxrrrrr staticmethodrr+r$rrrFrFysU),%g % :r "&!PX6&PrrFc#"^RIHpHpHp^RIHp^RIHpV!4'gRxR#V!V!\V!444\V!44R7pRxV!V4R# T!T4i;i5i)ajInstall the gateway launch profile's secret scope while multiplexing is on. The tick runs via ``_to_thread_process_service`` in a fresh context, so no per-turn scope exists and ``get_secret`` fails closed. The decomposer's aux LLM reads ``auxiliary.*`` from ``get_hermes_home()``, so its credentials come from that same home. No-op for single-profile gateways. )is_multiplex_activereset_secret_scopeset_secret_scope)get_hermes_home)owner_secret_scopeN) profile_home) agent.secret_scoperrrhermes_constantsr!tui_gateway.launch_profile_policyrrrW)rrrrrtokens rrr2s`]\0D    /_5F0GHWZ[j[lWm nE" 5!5!sAA4A' A4' A11A4c V^8dQhRRRR/#)r-resultszOptional[list]r1r r$)r2s"rr3r3Is4rcRpT;'g.Fwr#VfK \VRR4'gK Rp\P!RT\VP4VP \ VPR4'd\VP4M^\ VPR4'd\VP4M^VP\ VPR4'd\VP4M^4K V#)zALog per-board spawn summaries; returns whether any board spawned.FNspawnedTzckanban dispatcher [%s]: spawned=%d reclaimed=%d crashed=%d timed_out=%d promoted=%d auto_blocked=%d__len__) getattrr r=rr reclaimedhasattrcrashed timed_outpromoted auto_blocked)r any_spawnedrVress& r_log_spawn_resultsrIsKmmm  ?wsIt<<K KKFc#++& $+CKK$C$CCKK &-cmmY&G&GCMM"Q )01A1A9)M)MC$$%ST  $ r)zfile is not a databasez database disk image is malformed)r) __future__rrrrtr~ dataclassesrrpathlibrtypingrrgateway.kanban_watchers_commonrr r rrrwrrDrFcontextmanagerrrr$rrrs#  ) VV U  / /  /GTvvr "",r