@pellux/goodvibes-agent 1.2.0 → 1.4.1

package/release/release-readiness.json

@@ -62,7 +62,7 @@
       "quality": {
         "capabilityCoverage": "Matches the release expectation that the product opens into the real operator workspace, with visible assistant-first Home lanes for setup, chat/model, connected browser cockpit/PWA handoff, project work, Personal Ops, research/docs, background supervision, and safety/recovery plus deeper setup, runtime posture, local context, channels, media, planning, and delegation categories.",
         "userAccess": "The user can open the workspace by default or through /agent, /home, and /operator, start from the assistant cockpit lanes including the browser cockpit handoff, then browse or search deeper workspace categories and actions directly.",
-        "modelAccess": "route action:\"plan|status\" plus lower-level agent_harness mode:\"route_decision\" choose the preferred visible Agent route for a plain user task, including alternatives, missing fields, confirmation boundary, workspace matches, and harness mode matches across setup, settings, Personal Ops, research, autonomy, execution, delegation, computer/browser, workspace, host, device, channel, security, Local Context, and Knowledge surfaces without executing tools; host/daemon health, doctor, readiness, service, and compatibility wording prefers host action:\"status\" before repair or lifecycle effects, normal settings/configuration wording prefers settings action:\"list\" before set/reset/import effects, model provider/local-cookbook/local-server-smoke/route-fit wording prefers models action:\"provider|local|smoke|route\" before credential, smoke, benchmark, or route-change effects, Personal Ops briefing/queue/fresh-read/connector setup wording prefers personal_ops action:\"briefing|queue|intake|lane\" before live provider reads or effects, direct reminder/schedule/cron/lifecycle wording prefers schedule action:\"list\" before confirmed schedule effects, command-shaped background work prefers execution action:\"processes\" plus first-class terminal/process routes, interactive PTY/stdin/sudo wording prefers execution action:\"process_capabilities\" before hidden process starts or credential effects, external memory-provider/backend/sync/import-export wording prefers memory action:\"provider\" or the external provider checklist before provider effects, browser-backed research runner wording prefers research action:\"runner\" readiness, visual research report rendering wording prefers research action:\"plan\" plus report artifacts before browser/PWA rendering is claimed, voice workflow/TTS-provider wording prefers device action:\"voice|provider\" before capture, playback, or picker effects, browser cockpit/PWA wording prefers computer action:\"browser\" before visible open handoffs, channel setup/triage/delivery-receipt/send wording prefers channels action:\"setup|triage|deliveries|channel\" before confirmed external delivery, permission posture/finding/blocked-action wording prefers security action:\"status|finding|explain\" before policy changes or risky work, support-bundle/session/audit evidence wording prefers support action:\"status|bundle\", sessions action:\"list|get\", and audit action:\"readiness|evidence|item|artifact\" before bundle export/import/share, session lifecycle, or audit drill-in effects, broader ongoing or watcher-like background work stays on autonomy intake, file undo/redo/recovery wording prefers execution action:\"recovery\", media generation wording prefers media provider readiness plus confirmed agent_media_generate saved artifacts, and screenshot, browser-navigation/control, screen-observation, and desktop-control wording prefers computer action:\"plan\" before live UI tools are considered; workspace action:\"status|actions|action|run|surfaces|surface|open|panels|panel|open_panel|shortcuts|keybindings|keybinding|run_keybinding|set_keybinding|reset_keybinding|commands|command|run_command|cli_commands|cli_command\" plus autonomy action:\"intake|queue|item|status\" plus computer action:\"status|plan|control|browser|setup|mcp|open_browser\" plus device action:\"status|capability|voice|provider|open_tts_provider|open_tts_voice\" plus models action:\"status|route|local|providers|provider|smoke\" plus personal_ops action:\"briefing|status|queue|intake|lane|read\" plus channels action:\"status|channel|setup|triage|deliveries\" plus schedule action:\"list|create|remind|edit|run|pause|resume|delete\" expose the same assistant-first cockpit, workspace map, connected browser cockpit/PWA route with service/web setup fallback, workspace category coverage, visible autonomy queue live records, model provider/local-cookbook/smoke/route-fit routes, Personal Ops briefing/queue/fresh-read routes, channel setup/triage/delivery history, schedule post-action nextRoutes, mobile/PWA controls, Agent onboarding marker status, unpublished browser/PWA receipt reporting, visible navigation routes, keyboard route controls, and command/action execution to the model without requiring harness mode memorization; lower-level agent_harness mode:\"summary\", mode:\"route_decision\", mode:\"workspace\", mode:\"workspace_categories\", mode:\"workspace_actions\", mode:\"workspace_action\", mode:\"run_workspace_action\", mode:\"autonomy_intake\", mode:\"autonomy_queue\", mode:\"autonomy_queue_item\", mode:\"agent_orchestration\", mode:\"agent_orchestration_agent\", mode:\"panels\", mode:\"panel\", mode:\"open_panel\", mode:\"ui_surfaces\", mode:\"ui_surface\", mode:\"open_ui_surface\", mode:\"shortcuts\", mode:\"keybindings\", mode:\"keybinding\", mode:\"run_keybinding\", mode:\"set_keybinding\", and mode:\"reset_keybinding\" remain detailed compatibility routes for route inspection and release verification, including managed multi-agent milestones, work-plan dispatch receipts, task/run diagnostics, bounded redacted output descriptors, automation telemetry, and compact modelRoute hints.",
+        "modelAccess": "route action:\"plan|status\" plus lower-level agent_harness mode:\"route_decision\" choose the preferred visible Agent route for a plain user task, including alternatives, missing fields, confirmation boundary, workspace matches, and harness mode matches across setup, settings, Personal Ops, research, autonomy, execution, delegation, computer/browser, workspace, host, device, channel, security, Local Context, and Knowledge surfaces without executing tools; host/daemon health, doctor, readiness, service, and compatibility wording prefers host action:\"status\" before repair or lifecycle effects, normal settings/configuration wording prefers settings action:\"list\" before set/reset/import effects, model provider/local-cookbook/local-server-smoke/route-fit wording prefers models action:\"provider|local|smoke|route\" before credential, smoke, benchmark, or route-change effects, Personal Ops briefing/queue/fresh-read/connector setup wording prefers personal_ops action:\"briefing|queue|intake|lane\" before live provider reads or effects, direct reminder/schedule/cron/lifecycle wording prefers schedule action:\"list\" before confirmed schedule effects, command-shaped background work prefers execution action:\"processes\" plus first-class terminal/process routes, interactive PTY/stdin/sudo wording prefers execution action:\"process_capabilities\" before hidden process starts or credential effects and surfaces certified daemon runtime records when published, external memory-provider/backend/sync/import-export wording prefers memory action:\"provider\" or the certified external provider checklist before provider effects, browser-backed research runner wording prefers research action:\"runner\" readiness, visual research report rendering wording prefers research action:\"plan\" plus report artifacts before browser/PWA rendering is claimed, voice workflow/TTS-provider wording prefers device action:\"voice|provider\" before capture, playback, or picker effects, browser cockpit/PWA wording prefers computer action:\"browser\" before visible open handoffs, channel setup/triage/delivery-receipt/send wording prefers channels action:\"setup|triage|deliveries|channel\" before confirmed external delivery, permission posture/finding/blocked-action wording prefers security action:\"status|finding|explain\" before policy changes or risky work, support-bundle/session/audit evidence wording prefers support action:\"status|bundle\", sessions action:\"list|get\", and audit action:\"readiness|evidence|item|artifact\" before bundle export/import/share, session lifecycle, or audit drill-in effects, broader ongoing or watcher-like background work stays on autonomy intake, file undo/redo/recovery wording prefers execution action:\"recovery\", media generation wording prefers media provider readiness plus confirmed agent_media_generate saved artifacts, and screenshot, browser-navigation/control, screen-observation, and desktop-control wording prefers computer action:\"plan\" before live UI tools are considered; workspace action:\"status|actions|action|run|surfaces|surface|open|panels|panel|open_panel|shortcuts|keybindings|keybinding|run_keybinding|set_keybinding|reset_keybinding|commands|command|run_command|cli_commands|cli_command\" plus autonomy action:\"intake|queue|item|status\" plus computer action:\"status|plan|control|browser|setup|mcp|open_browser\" plus device action:\"status|capability|voice|provider|open_tts_provider|open_tts_voice\" plus models action:\"status|route|local|providers|provider|smoke\" plus personal_ops action:\"briefing|status|queue|intake|lane|read\" plus channels action:\"status|channel|setup|triage|deliveries\" plus schedule action:\"list|create|remind|edit|run|pause|resume|delete\" expose the same assistant-first cockpit, workspace map, connected browser cockpit/PWA route with service/web setup fallback, workspace category coverage, visible autonomy queue live records, model provider/local-cookbook/smoke/route-fit routes, Personal Ops briefing/queue/fresh-read routes, channel setup/triage/delivery history, schedule post-action nextRoutes, mobile/PWA controls, Agent onboarding marker status, browser/PWA receipt reporting that distinguishes satisfied artifact/live read-model evidence from missing publication, visible navigation routes, keyboard route controls, and command/action execution to the model without requiring harness mode memorization; lower-level agent_harness mode:\"summary\", mode:\"route_decision\", mode:\"workspace\", mode:\"workspace_categories\", mode:\"workspace_actions\", mode:\"workspace_action\", mode:\"run_workspace_action\", mode:\"autonomy_intake\", mode:\"autonomy_queue\", mode:\"autonomy_queue_item\", mode:\"agent_orchestration\", mode:\"agent_orchestration_agent\", mode:\"panels\", mode:\"panel\", mode:\"open_panel\", mode:\"ui_surfaces\", mode:\"ui_surface\", mode:\"open_ui_surface\", mode:\"shortcuts\", mode:\"keybindings\", mode:\"keybinding\", mode:\"run_keybinding\", mode:\"set_keybinding\", and mode:\"reset_keybinding\" remain detailed compatibility routes for route inspection and release verification, including managed multi-agent milestones, work-plan dispatch receipts, task/run diagnostics, bounded redacted output descriptors, automation telemetry, and compact modelRoute hints.",
         "safetyBoundary": "The workspace stays Agent-first and avoids hidden connected-host lifecycle control, coding-first entrypoints, and marketing-only surfaces.",
         "releaseEvidence": "release-readiness inventory plus package-facing README and getting-started docs cite the workspace files and keep this as a required release item."
       }
@@ -72,14 +72,14 @@
       "capability": "First-run setup covers provider access, Agent Knowledge, local behavior libraries, channels, voice/media, and connected-host posture.",
       "owner": "agent",
       "status": "covered",
-      "evidence": "src/agent/setup-wizard-checkpoint.ts; src/input/agent-workspace.ts; src/input/agent-workspace-categories.ts; src/runtime/onboarding/verify.ts; src/tools/agent-settings-import-tool.ts; src/tools/agent-harness-setup-posture.ts",
+      "evidence": "src/agent/setup-wizard-checkpoint.ts; src/agent/setup-wizard-artifact-receipts.ts; src/input/setup-wizard-live-receipts.ts; src/input/agent-workspace.ts; src/input/agent-workspace-categories.ts; src/runtime/onboarding/verify.ts; src/tools/agent-settings-import-tool.ts; src/tools/agent-harness-setup-posture.ts",
       "action": "Keep setup review tied to the Agent workspace, Agent-owned settings, and connected-host readiness checks.",
       "quality": {
-        "capabilityCoverage": "Covers first-run provider setup, a guided setup wizard with progress/current-step/saved-checkpoint/backtracking/repeated-smoke-blocker focus, visible connected-host auth and install-smoke checklist rows, GoodVibes settings import, Agent Knowledge readiness, VIBE.md personality health, local behavior libraries, channels, voice/media, browser/desktop control readiness, sudo execution posture with SUDO_PASSWORD presence-only guidance, support posture, connected-host compatibility, live service probe evidence, token-safe connected-host auth posture with exact pairing route ids, confirmed SDK-backed local token create/repair, token-safe install smoke checks, confirmed setup smoke execution, optional durable redacted setup smoke artifacts, latest-smoke and history/trend Home/setup summary surfacing, primary handoff cards for every actionable setup row, read-only setup repair decisions for the safest next token/status/bootstrap/lifecycle/no-action route, setup closeout decisions from critical blocker, saved smoke, and completion-marker evidence, confirmed onboarding finish through the workspace action route, offline GoodVibes host bootstrap guidance, recommended diagnostic/status repair cards, setup-state fixtures, setup serviceLifecycleDecision gates, inspect-first confirmed connected-host service repair routes through the Agent workspace, certified service receipt outcomes for services.status/install/start/restart, and exact install/start/restart/no-action lifecycle decisions from services.status receipts.",
-        "userAccess": "The user gets the Agent workspace with setup readiness, a guided Start wizard that highlights the current setup step, saved setup checkpoint, stale-checkpoint auto-advance evidence, repeated smoke blocker, and closeout state, visible Start actions to show/save/clear the setup checkpoint, visible connected-host auth and install-smoke checklist rows beside runtime/model setup, current settings, GoodVibes import preview/apply, reviewable readiness state, primary handoffs from setup blockers to the safest visible workspace form, diagnostic, user-run guidance, or confirmed route, sudo posture with blocked background escalation routes and presence-only credential status, a direct setup repair decision route for the current blocker or a named host/auth/service target, confirmed local connected-host token create/repair, confirmed token-safe setup smoke, saved redacted setup smoke evidence artifacts with latest result, trend, recent runs, and frequent blockers surfaced in the setup summary, token-safe connected-host auth status with exact pairing route ids, setup closeout that says fix blocker, run final smoke, finish setup, or complete, confirmed Apply & close marker write, offline install/trust/verify/service/reconnect commands when no host is reachable, exact service status/install/start/restart repair options when the connected-host operator endpoint supports them, serviceLifecycleDecision status gates, and success criteria plus certified receipt outcomes and exact lifecycle decisions when service methods return status proof.",
-        "modelAccess": "setup action:\"status|item|repair|checkpoint|save_checkpoint|clear_checkpoint|token|smoke|finish|import_settings\" gives the preferred direct first-run setup route; settings action:\"list|get|set|reset|import\" gives the preferred direct Agent settings and redacted shared GoodVibes settings import route; host action:\"status|capabilities|capability|services|service|methods|method\" gives the preferred direct connected-host diagnostics route; import_goodvibes_settings action:\"preview|apply\" remains available for compatibility; agent_harness mode:\"summary\", mode:\"setup_posture\", mode:\"setup_item\", mode:\"setup_repair\", mode:\"setup_checkpoint\", mode:\"mark_setup_checkpoint\", mode:\"clear_setup_checkpoint\", mode:\"provision_connected_host_token\", mode:\"run_setup_smoke\", mode:\"settings\", mode:\"get_setting\", mode:\"set_setting\", and mode:\"reset_setting\" remain available for detailed setup/settings compatibility and expose setup marker state, a guided setupWizard with progress/current step/checkpoint state/checkpoint autoAdvance evidence/backtracking routes/rerun and save-smoke routes/repeated-blocker focus, setupWizard.closeout and top-level setupCloseout decisions with blocker/smoke/marker evidence plus the exact finish route, a prioritized first-run setup plan with connected-host/provider/auth blockers, sudo execution posture, primaryHandoff and handoffs cards with kind/effect/userRoute/modelRoute/safety/confirmation metadata, nextSetupActions and summary nextSetupHandoffs handoff routes, a read-only setup repair decision that selects token repair, connected-host status, services.status receipt, user-run bootstrap, or no lifecycle action without executing it, confirmed setup checkpoint save/clear, confirmed local token provisioning, token-safe install smoke checks, confirmed redacted smoke evidence collection, optional artifact capture for user-run smoke output, latest saved smoke artifact result, setup smoke history/trend/recent-run/frequent-blocker metadata, local model readiness and cookbook follow-through, VIBE.md personality health counts for applied/blocked/truncated files, shared GoodVibes settings import counts, browser/desktop-control ready-attention-setup readiness, SUDO_PASSWORD presence-only guidance with blocked background sudo/stdin password routes, exact follow-up routes, live service probe evidence, setup serviceLifecycleDecision gates, token-safe connected-host auth posture with pairing route ids, recommended repair cards, offline GoodVibes host bootstrap commands, connected-host repair cards with success criteria, certified service repair receipt outcomes and lifecycleDecision outputs from agent_operator_method, derived capability flags, provider route, provider-account counts, subscription counts, local behavior discovery, channel/media/setup signals, single setup-item lookup, and checkpoint inspection; mode:\"run_workspace_action\" actionId:\"onboarding-apply-close\" writes the confirmed user onboarding markers, mode:\"run_workspace_action\" actionId:\"import-goodvibes-tui-settings\" remains available for visible workspace parity; settings rows include compact first-class modelRoute hints, and visible UI, workspace, and command routes remain for confirmed setup mutations.",
-        "safetyBoundary": "Setup reads and writes Agent-owned settings only, keeps token values redacted, imports secret-backed values through the secret manager, stores setup checkpoints as only the visible setup step id, label, source, timestamp, and generic note without explicit user request text, exposes setup repair as a read-only decision that never starts, installs, restarts, writes tokens, imports settings, or opens UI by itself, exposes missing-host bootstrap commands as user-run guidance, runs setup smoke only as a confirmed redacted evidence collection route without implicit shell/package/host commands, redacts token/secret/password/credential/authorization/api-key-like values before optional artifact storage, treats setup closeout as advisory until confirmed Apply & close writes only the user onboarding markers, exposes sudo as credential presence, blocked route, and foreground guidance only without reading, storing, printing, or injecting raw password values, exposes connected-host auth as presence/path/fingerprint posture plus pairing route ids only, creates or repairs the local canonical operator token only through confirmed SDK-backed setup while preserving valid tokens and leaving env tokens untouched, exposes connected-host service install/start/restart only as confirmed operator-method repair routes that remain inspect-first until services.status lifecycleDecision evidence proves need, and intentionally excludes service stop/uninstall from first-run setup.",
-        "releaseEvidence": "Readiness evidence cites Agent workspace, verification, and direct harness sources, while ledger accounting and package-facing docs cover model-visible setup posture."
+        "capabilityCoverage": "Covers first-run provider setup, a guided setup wizard with progress/current-step/saved-checkpoint/backtracking/repeated-smoke-blocker focus, visible connected-host auth and install-smoke checklist rows, GoodVibes settings import, Agent Knowledge readiness, VIBE.md personality health, local behavior libraries, channels, voice/media, browser/desktop control readiness, sudo execution posture with SUDO_PASSWORD presence-only guidance, support posture, connected-host compatibility, live service probe evidence, token-safe connected-host auth posture with exact pairing route ids, confirmed SDK-backed local token create/repair, token-safe install smoke checks, confirmed setup smoke execution, optional durable redacted setup smoke artifacts, durable setup receipts from saved artifacts, live setup read-model snapshots, and ordered setup receipt event streams, certified setup receipt schema/version/provenance/publication/cursor evidence, latest-smoke and history/trend Home/setup summary surfacing, primary handoff cards for every actionable setup row, read-only setup repair decisions for the safest next token/status/bootstrap/lifecycle/no-action route, setup closeout decisions from critical blocker, saved smoke, durable setup receipt, and completion-marker evidence, confirmed onboarding finish through the workspace action route, offline GoodVibes host bootstrap guidance, recommended diagnostic/status repair cards, setup-state fixtures, setup serviceLifecycleDecision gates, inspect-first confirmed connected-host service repair routes through the Agent workspace, certified service receipt outcomes for services.status/install/start/restart, and exact install/start/restart/no-action lifecycle decisions from services.status receipts.",
+        "userAccess": "The user gets the Agent workspace with setup readiness, a guided Start wizard that highlights the current setup step, saved setup checkpoint, stale-checkpoint auto-advance evidence, repeated smoke blocker, and closeout state, visible Start actions to show/save/clear the setup checkpoint, visible connected-host auth and install-smoke checklist rows beside runtime/model setup, current settings, GoodVibes import preview/apply, reviewable readiness state, primary handoffs from setup blockers to the safest visible workspace form, diagnostic, user-run guidance, or confirmed route, sudo posture with blocked background escalation routes and presence-only credential status, a direct setup repair decision route for the current blocker or a named host/auth/service target, confirmed local connected-host token create/repair, confirmed token-safe setup smoke, saved redacted setup smoke evidence artifacts, live/saved durable setup receipt auto-advance for service/auth/install-smoke/browser-PWA rows, ordered event-stream setup receipts with schema/provenance/publication/cursor evidence, and latest result, trend, recent runs, and frequent blockers surfaced in the setup summary, token-safe connected-host auth status with exact pairing route ids, setup closeout that says fix blocker, run final smoke, finish setup, or complete, confirmed Apply & close marker write, offline install/trust/verify/service/reconnect commands when no host is reachable, exact service status/install/start/restart repair options when the connected-host operator endpoint supports them, serviceLifecycleDecision status gates, and success criteria plus certified receipt outcomes and exact lifecycle decisions when service methods return status proof.",
+        "modelAccess": "setup action:\"status|item|repair|checkpoint|save_checkpoint|clear_checkpoint|token|smoke|finish|import_settings\" gives the preferred direct first-run setup route; settings action:\"list|get|set|reset|import\" gives the preferred direct Agent settings and redacted shared GoodVibes settings import route; host action:\"status|capabilities|capability|services|service|methods|method\" gives the preferred direct connected-host diagnostics route; import_goodvibes_settings action:\"preview|apply\" remains available for compatibility; agent_harness mode:\"summary\", mode:\"setup_posture\", mode:\"setup_item\", mode:\"setup_repair\", mode:\"setup_checkpoint\", mode:\"mark_setup_checkpoint\", mode:\"clear_setup_checkpoint\", mode:\"provision_connected_host_token\", mode:\"run_setup_smoke\", mode:\"settings\", mode:\"get_setting\", mode:\"set_setting\", and mode:\"reset_setting\" remain available for detailed setup/settings compatibility and expose setup marker state, a guided setupWizard with progress/current step/checkpoint state/checkpoint autoAdvance evidence/backtracking routes/rerun and save-smoke routes/repeated-blocker focus, setupWizard.closeout and top-level setupCloseout decisions with blocker/smoke/marker evidence plus certified receipt counts and event-stream counts, a prioritized first-run setup plan with connected-host/provider/auth blockers, sudo execution posture, primaryHandoff and handoffs cards with kind/effect/userRoute/modelRoute/safety/confirmation metadata, nextSetupActions and summary nextSetupHandoffs handoff routes, a read-only setup repair decision that selects token repair, connected-host status, services.status receipt, user-run bootstrap, or no lifecycle action without executing it, confirmed setup checkpoint save/clear, confirmed local token provisioning, token-safe install smoke checks, confirmed redacted smoke evidence collection, optional artifact capture for user-run smoke output, latest saved smoke artifact result, merged durable setup receipt artifacts, live setup read-model snapshots, and ordered setup receipt event streams for service/auth/install-smoke/browser-PWA auto-advance with redacted summaries, source ids, schema status/version, repair/action provenance, host publication guarantees, publisher, cursor, and sequence evidence, setup smoke history/trend/recent-run/frequent-blocker metadata, local model readiness and cookbook follow-through, VIBE.md personality health counts for applied/blocked/truncated files, shared GoodVibes settings import counts, browser/desktop-control ready-attention-setup readiness, SUDO_PASSWORD presence-only guidance with blocked background sudo/stdin password routes, exact follow-up routes, live service probe evidence, setup serviceLifecycleDecision gates, token-safe connected-host auth posture with pairing route ids, recommended repair cards, offline GoodVibes host bootstrap commands, connected-host repair cards with success criteria, certified service repair receipt outcomes and lifecycleDecision outputs from agent_operator_method, derived capability flags, provider route, provider-account counts, subscription counts, local behavior discovery, channel/media/setup signals, single setup-item lookup, and checkpoint inspection; mode:\"run_workspace_action\" actionId:\"onboarding-apply-close\" writes the confirmed user onboarding markers, mode:\"run_workspace_action\" actionId:\"import-goodvibes-tui-settings\" remains available for visible workspace parity; settings rows include compact first-class modelRoute hints, and visible UI, workspace, and command routes remain for confirmed setup mutations.",
+        "safetyBoundary": "Setup reads and writes Agent-owned settings only, keeps token values redacted, imports secret-backed values through the secret manager, stores setup checkpoints as only the visible setup step id, label, source, timestamp, and generic note without explicit user request text, exposes setup repair as a read-only decision that never starts, installs, restarts, writes tokens, imports settings, or opens UI by itself, exposes missing-host bootstrap commands as user-run guidance, runs setup smoke only as a confirmed redacted evidence collection route without implicit shell/package/host commands, redacts token/secret/password/credential/authorization/api-key-like values before optional artifact storage or setup receipt history/detail surfacing, treats setup closeout as advisory until confirmed Apply & close writes only the user onboarding markers, exposes sudo as credential presence, blocked route, and foreground guidance only without reading, storing, printing, or injecting raw password values, exposes connected-host auth as presence/path/fingerprint posture plus pairing route ids only, creates or repairs the local canonical operator token only through confirmed SDK-backed setup while preserving valid tokens and leaving env tokens untouched, exposes connected-host service install/start/restart only as confirmed operator-method repair routes that remain inspect-first until services.status lifecycleDecision evidence proves need, and intentionally excludes service stop/uninstall from first-run setup.",
+        "releaseEvidence": "Readiness evidence cites Agent workspace, verification, and direct harness sources, while ledger accounting and package-facing docs cover model-visible setup posture plus artifact, live read-model, and ordered event-stream setup receipt consumption."
       }
     },
     {
@@ -87,14 +87,14 @@
       "capability": "Provider and model selection includes active model visibility, custom provider setup, model refresh, pinned models, and reasoning effort.",
       "owner": "agent",
       "status": "covered",
-      "evidence": "src/input/agent-workspace-categories.ts; src/input/commands/local-provider-runtime.ts; src/input/commands/shell-core.ts; src/tools/agent-harness-model-routing.ts; docs/providers-and-routing.md",
+      "evidence": "src/input/agent-workspace-categories.ts; src/input/commands/local-provider-runtime.ts; src/input/commands/shell-core.ts; src/tools/agent-harness-model-routing.ts; src/tools/agent-harness-local-model-endpoints.ts; docs/providers-and-routing.md",
       "action": "Preserve workspace-first model/provider controls and scriptable mirrors.",
       "quality": {
-        "capabilityCoverage": "Covers active model visibility, custom provider setup, model refresh, favorites, pinned models, context windows, provider-health posture split across SDK type availability, daemon publication, and Agent consumption, reasoning effort controls, visible route-readiness inspection, readiness scoring across latency/context/tools/vision/cost/privacy, hardware-scored local model recommendations, setup/download guidance, local endpoint candidates with exact endpoint inspection, model-list smoke commands, success criteria, failure triage, confirmed models action:\"smoke\" checks, provider-add and refresh route hints, confirmed local benchmark execution through agent_model_compare, saved benchmark-evidence review, saved local benchmark history with filtered benchmark analytics, and revealed winner judgments feeding matching recipe confidence.",
-        "userAccess": "The user can operate provider and model routing from the Agent workspace, /model, /provider, model picker, provider workspace, scriptable CLI mirrors, and explicit Model Routing actions for route readiness, local cookbook guidance, confirmed local server checks, local benchmark execution, and benchmark evidence review.",
-        "modelAccess": "Route planning prefers models action:\"provider\" for provider account/subscription wording, action:\"local\" for Ollama/local cookbook/hardware-fit wording, action:\"smoke\" for local server health checks, and action:\"route\" for context/tool/latency/cost/privacy route-fit wording before credential, smoke, benchmark, or route-change effects; models action:\"status|route|local|providers|provider|smoke\" exposes current provider/model route, selectable models, provider ids, provider/subscription posture, pinned models, reasoning support, context windows, readiness scores/dimensions, providerHealth with SDK provider-health type availability, daemon publication status, Agent consumption status, missing signals, and live latency when a daemon-published read model is reachable, safe route setting keys, hardware-scored local model cookbook, setup/download guidance, local endpoint candidates with exact route/provider inspection, model-list smoke commands, success criteria, failure triage, confirmed local model-list smoke results, provider-add and refresh route hints, confirmed local benchmark workspace action route, saved local benchmark history with judgment evidence and a benchmark-filtered analytics route, and single route/model/endpoint/provider lookup; lower-level agent_harness mode:\"model_routing\", mode:\"model_route\", mode:\"provider_accounts\", mode:\"provider_account\", and confirmed mode:\"run_local_model_smoke\" remain compatibility routes; workspace action hints expose Inspect route readiness, Check local servers through models action:\"smoke\", and Review benchmark evidence; visible picker, settings, workspace, and command routes remain for confirmed selection, refresh, benchmark execution, judgment review/apply, pin/unpin, custom provider, and setting mutations.",
-        "safetyBoundary": "Provider auth and model changes are explicit, secret-backed where needed, and do not print raw credentials.",
-        "releaseEvidence": "Provider routing docs, direct harness source, package-facing text checks, and readiness item evidence keep this feature required for stable releases."
+        "capabilityCoverage": "Covers active model visibility, custom provider setup, model refresh, favorites, pinned models, context windows, provider-health posture split across SDK type availability, daemon publication, Agent consumption, provider fallback, exact route-health read models, live latency, rate-limit posture, and redacted error posture, reasoning effort controls, visible route-readiness inspection, readiness scoring across latency/context/tools/vision/cost/privacy, hardware-scored local model recommendations, setup/download guidance, local endpoint candidates with exact endpoint inspection, daemon-published local serving diagnostics for certified schema/version/provenance/publication/publisher metadata, server version, loaded models, context/tool support, resource pressure, start/repair receipt ids, receipt status, exact confirmed start/repair routes when host-published, and redacted summaries, model-list smoke commands, success criteria, failure triage, confirmed models action:\"smoke\" checks, provider-add and refresh route hints, confirmed local benchmark execution through agent_model_compare, saved benchmark-evidence review, saved local benchmark history with filtered benchmark analytics, and revealed winner judgments feeding matching recipe confidence.",
+        "userAccess": "The user can operate provider and model routing from the Agent workspace, /model, /provider, model picker, provider workspace, scriptable CLI mirrors, and explicit Model Routing actions for route readiness, local cookbook guidance, route-level provider health, certified daemon-published local serving diagnostics when available, confirmed local server checks, local benchmark execution, and benchmark evidence review.",
+        "modelAccess": "Route planning prefers models action:\"provider\" for provider account/subscription wording, action:\"local\" for Ollama/local cookbook/hardware-fit wording, action:\"smoke\" for local server health checks, and action:\"route\" for context/tool/latency/cost/privacy route-fit wording before credential, smoke, benchmark, or route-change effects; models action:\"status|route|local|providers|provider|smoke\" exposes current provider/model route, selectable models, provider ids, provider/subscription posture, pinned models, reasoning support, context windows, readiness scores/dimensions, providerHealth with SDK provider-health type availability, daemon publication status, Agent consumption status, exact route-health matching, source record ids, missing signals, live latency, rate-limit posture, and redacted error posture when a daemon-published read model is reachable, safe route setting keys, hardware-scored local model cookbook, setup/download guidance, local endpoint candidates with exact route/provider inspection, daemonDiagnostics publication posture, endpoint servingDiagnostics for certified schema/version/provenance/publication/publisher metadata, server version, loaded models, context/tool support, resource pressure, start/repair receipt ids, receipt status, source path, exact confirmed start/repair routes when host-published, redacted summaries, model-list smoke commands, success criteria, failure triage, confirmed local model-list smoke results, provider-add and refresh route hints, confirmed local benchmark workspace action route, saved local benchmark history with judgment evidence and a benchmark-filtered analytics route, and single route/model/endpoint/provider lookup; lower-level agent_harness mode:\"model_routing\", mode:\"model_route\", mode:\"provider_accounts\", mode:\"provider_account\", and confirmed mode:\"run_local_model_smoke\" remain compatibility routes; workspace action hints expose Inspect route readiness, Check local servers through models action:\"smoke\", and Review benchmark evidence; visible picker, settings, workspace, and command routes remain for confirmed selection, refresh, benchmark execution, judgment review/apply, pin/unpin, custom provider, and setting mutations.",
+        "safetyBoundary": "Provider auth and model changes are explicit, secret-backed where needed, and do not print raw credentials. Daemon-published provider health and local serving diagnostics are read-only metadata consumption with secret redaction; host-published start/repair routes are only surfaced as separate confirmed actions, and Agent still requires separate confirmation for local smoke probes, provider edits, catalog refresh, benchmark execution, server start/repair, and route changes.",
+        "releaseEvidence": "Provider routing docs, direct harness source, local serving diagnostics tests, package-facing text checks, and readiness item evidence keep this feature required for stable releases."
       }
     },
     {
@@ -137,7 +137,7 @@
       "quality": {
         "capabilityCoverage": "Covers explicit handling for artifacts, files, URL lists, bookmarks, browser history, generated media, reviewed source promotion, document-operation readiness, visible review packet timeline including stale preset attention, visible review packet wizard progress/routes/refreshed-preset lineage, reusable review packet preset artifacts with missing/superseded id freshness checks and confirmed refresh-to-new-preset routes, confirmed review packet archive-reference sharing, route-decision receipt artifacts for apply and leave-unchanged evidence, visible reviewer-readiness preflight, review packet defaults that prefill export/handoff/archive/apply/leave-unchanged/save-preset/share forms from latest packet evidence with preset fallback only when live evidence is missing, inline readiness badges before export/archive/apply, reviewer-ready document export appendices, confirmed artifact package directory or ZIP archive export, and confirmed blind model comparison side-by-side, handoff diff section jumps with recent-handoff choices, plus task/document/benchmark-filtered analytics/synthesis/handoff/handoff archive with matching route-decision receipt evidence.",
         "userAccess": "The user can use workspace ingest forms, reviewer-ready document export, review packet share, artifact export/package/archive forms, media generation, notes, bookmarks, and Knowledge ingest mirrors with clear review steps.",
-        "modelAccess": "agent_knowledge_ingest exposes URL/file/URL-list/bookmarks/browser-history/connector ingest, agent_documents exposes confirmed document artifact export with comment/suggestion summary appendices and review metadata counts, agent_review_packet_presets exposes confirmed save and refresh plus read-only list/show freshness checks and recommended reuse routes for reusable review packet preset artifacts without changing documents, model routing, handoffs, archives, or source presets, agent_review_packet_share validates saved reviewer handoff archives and sends confirmed plain-text archive references through configured channel targets without printing or attaching ZIP bytes, agent_artifacts exposes saved artifact browse/show/export/package/archive routes with confirmation-gated writes, research exposes a read-only next-action briefing, read-only planning, bounded public source-candidate search with confirmed capture routes, direct browser-runner readiness, visible project-local run/source inspection, reviewed-source bundles, saved report artifact inspection, confirmed run lifecycle controls, confirmed source triage, and sourced report artifact saves with citation coverage repair metadata, optional strict body-citation enforcement, and optional visual report packet sections; lower-level agent_research_runs, agent_research_sources, agent_research_report, agent_harness mode:\"research_briefing\", agent_harness mode:\"research_workflow\", agent_harness mode:\"research_runs\", agent_harness mode:\"research_run\", agent_harness mode:\"research_queue\", and agent_harness mode:\"research_source\" remain detail routes; Agent Workspace Research exposes browser-runner and visual-report readiness plus Research briefing, Plan workflow, public search, direct runner, and report artifact actions for research action:\"briefing|plan|search|runner|reports\", device exposes direct device/voice/browser/PWA/provider posture and confirmed visible browser/TTS picker handoffs, agent_media_generate exposes confirmed generated artifacts, agent_model_compare exposes confirmed blind model comparison with delayed reveal plus side-by-side reviewer views, section-jump reviewer handoff diffs with recent-handoff choices from artifact metadata, task/document/benchmark-filtered saved judgment analytics/synthesis, apply-winner and leave-unchanged route-decision receipts, reviewer handoff artifacts, and reviewer handoff ZIP archives that auto-include matching route-decision receipt artifacts in README counts, archive metadata, and redacted manifests, research action:\"briefing\" exposes one read-only next-action queue across visible runs, source review, saved report artifacts, browser readiness, and exact follow-up routes; research action:\"runner\" exposes the direct browser-runner readiness contract; research action:\"plan\" exposes a read-only ordered route plan across run state, public web/fetch or browser posture, browser-runner contracts, source capture/review, visual-report packet saving, and optional Knowledge promotion; research action:\"search\" exposes bounded public web result packets, source summaries, redacted capture arguments, and exact confirmed add_source routes without writing the queue; research action:\"runs|run\" exposes research run phase/progress/log-tail/control posture plus browser-backed runner readiness, fallback routes, source review, report, and Knowledge promotion routes; research action:\"sources|source|bundle\" exposes source credibility, bundle, and report handoff posture; research action:\"reports|report_artifact\" exposes saved report artifact lists, visual packet metadata, citation coverage, and bounded previews, mode:\"document_ops\" and mode:\"document_ops_lane\" expose documents, review packet timeline state including stale preset attention, review packet wizard progress/routes/refreshed-preset lineage/share handoff, route-decision receipt state, review packet preset/default freshness state, visible reviewer-readiness preflight, inline readiness badge state in workspace editor snapshots before export/archive/apply, uploads, exports, source checks, media artifacts, artifact browsing, and model comparison readiness, device action:\"voice|provider\" exposes media readiness while mode:\"media_posture\" and mode:\"media_provider\" remain detail routes, and mode:\"sessions\" and mode:\"session\" exposes bookmark/session artifact posture.",
+        "modelAccess": "agent_knowledge_ingest exposes URL/file/URL-list/bookmarks/browser-history/connector ingest, agent_documents exposes confirmed document artifact export with comment/suggestion summary appendices and review metadata counts, agent_review_packet_presets exposes confirmed save and refresh plus read-only list/show freshness checks and recommended reuse routes for reusable review packet preset artifacts without changing documents, model routing, handoffs, archives, or source presets, agent_review_packet_share validates saved reviewer handoff archives and sends confirmed plain-text archive references through configured channel targets without printing or attaching ZIP bytes, agent_artifacts exposes saved artifact browse/show/export/package/archive routes with confirmation-gated writes, research exposes a read-only next-action briefing, read-only planning, bounded public source-candidate search with confirmed capture routes, direct browser-runner readiness, certified SDK/daemon browser-backed runner records with schema/version/publication/publisher/provenance/freshness-cursor/receipt metadata, redacted current URLs/log tails, visible control routes, source/page receipt ids, certified browser/PWA visual report render records with render routes, redacted render URLs, visual sections, citation coverage, and source-map evidence, visible project-local run/source inspection, reviewed-source bundles, saved report artifact inspection, confirmed run lifecycle controls, confirmed source triage, and sourced report artifact saves with citation coverage repair metadata, optional strict body-citation enforcement, and optional visual report packet sections; lower-level agent_research_runs, agent_research_sources, agent_research_report, agent_harness mode:\"research_briefing\", agent_harness mode:\"research_workflow\", agent_harness mode:\"research_runs\", agent_harness mode:\"research_run\", agent_harness mode:\"research_queue\", and agent_harness mode:\"research_source\" remain detail routes; Agent Workspace Research exposes browser-runner and visual-report readiness plus Research briefing, Plan workflow, public search, direct runner, and report artifact actions for research action:\"briefing|plan|search|runner|reports\", device exposes direct device/voice/browser/PWA/provider posture and confirmed visible browser/TTS picker handoffs, agent_media_generate exposes confirmed generated artifacts, agent_model_compare exposes confirmed blind model comparison with delayed reveal plus side-by-side reviewer views, section-jump reviewer handoff diffs with recent-handoff choices from artifact metadata, task/document/benchmark-filtered saved judgment analytics/synthesis, apply-winner and leave-unchanged route-decision receipts, reviewer handoff artifacts, and reviewer handoff ZIP archives that auto-include matching route-decision receipt artifacts in README counts, archive metadata, and redacted manifests, research action:\"briefing\" exposes one read-only next-action queue across visible runs, source review, saved report artifacts, certified live runner/render evidence, browser readiness, and exact follow-up routes; research action:\"runner\" exposes the direct browser-runner readiness contract; research action:\"plan\" exposes a read-only ordered route plan across run state, public web/fetch or browser posture, browser-runner contracts, certified live browser-runner records, certified visual-render records, source capture/review, visual-report packet saving, and optional Knowledge promotion; research action:\"search\" exposes bounded public web result packets, source summaries, redacted capture arguments, and exact confirmed add_source routes without writing the queue; research action:\"runs|run\" exposes research run phase/progress/log-tail/control posture plus browser-backed runner readiness, certified live runner/render evidence, fallback routes, source review, report, and Knowledge promotion routes; research action:\"sources|source|bundle\" exposes source credibility, bundle, and report handoff posture; research action:\"reports|report_artifact\" exposes saved report artifact lists, visual packet metadata, citation coverage, certified renderer evidence, and bounded previews, mode:\"document_ops\" and mode:\"document_ops_lane\" expose documents, review packet timeline state including stale preset attention, review packet wizard progress/routes/refreshed-preset lineage/share handoff, route-decision receipt state, review packet preset/default freshness state, visible reviewer-readiness preflight, inline readiness badge state in workspace editor snapshots before export/archive/apply, uploads, exports, source checks, media artifacts, artifact browsing, and model comparison readiness, device action:\"voice|provider\" exposes media readiness while mode:\"media_posture\" and mode:\"media_provider\" remain detail routes, and mode:\"sessions\" and mode:\"session\" exposes bookmark/session artifact posture.",
         "safetyBoundary": "No artifact, generated media, research run, or research source is silently promoted into durable memory or non-Agent knowledge.",
         "releaseEvidence": "Readiness and docs cite Knowledge/artifact/media boundaries, and package policy requires those docs in the installable package."
       }
@@ -150,9 +150,9 @@
       "evidence": "src/agent/vibe-file.ts; src/agent/vibe-confirmation-routes.ts; src/tools/agent-vibe-tool.ts; src/tools/agent-personal-ops-tool.ts; src/tools/agent-harness-personal-ops.ts; src/agent/project-context-files.ts; src/agent/prompt-context-receipts.ts; src/tools/agent-harness-project-context.ts; src/tools/agent-harness-prompt-context.ts; src/tools/agent-harness-learning-curator.ts; src/tools/agent-learning-consolidation-tool.ts; src/input/agent-workspace-snapshot.ts; src/renderer/agent-workspace.ts; src/input/commands/vibe-runtime.ts; src/tools/agent-harness-memory-posture.ts; src/agent/persona-registry.ts; src/agent/note-registry.ts; src/cli/memory-command.ts; src/input/commands/personas-runtime.ts",
       "action": "Keep local context libraries Agent-owned and secret-rejecting.",
       "quality": {
-        "capabilityCoverage": "Covers safe project/global VIBE.md personality discovery, direct vibe status/show/init/import-persona model routes, VIBE init/import preview confirmationRoutes, safe project context discovery for .hermes.md, HERMES.md, AGENTS.md, CLAUDE.md, HERMES_HOME/SOUL.md, .cursorrules, and .cursor/rules/*.mdc files, visible Local Context and Personas workspace health counts for VIBE.md/project context files, Local Context prompt receipt outcome timeline with exact drill-in and outcome filters, status/init/show/import-persona flows, setup and learning-curator health cards for blocked/truncated VIBE.md files, opt-in VIBE.md profile starter export/import/application, durable local memory, memory posture, vector/embedding health, scratch notes, personas, profiles, search/review/stale/delete flows, duplicate-consolidation preview/merge/stale/delete/rollback/recreate receipts with exact-id post-delete recreate guidance, prompt injection gates for VIBE.md, project context files, and reviewed/confident setup-ready behavior, read-only prompt-context inspection plus sanitized prompt-context receipts for applied order/recent receipt ids/exact receipt-turn-outcome filters/selected records/suppressed records/token budget/prompt hash/terminal outcome, score-driven learning prompt plans for prompt-active/suppressed/proposal/consolidation queues, and promotion without knowledge contamination.",
+        "capabilityCoverage": "Covers safe project/global VIBE.md personality discovery, direct vibe status/show/init/import-persona model routes, VIBE init/import preview confirmationRoutes, safe project context discovery for .hermes.md, HERMES.md, AGENTS.md, CLAUDE.md, HERMES_HOME/SOUL.md, .cursorrules, and .cursor/rules/*.mdc files, visible Local Context and Personas workspace health counts for VIBE.md/project context files, Local Context prompt receipt outcome timeline with exact drill-in and outcome filters, status/init/show/import-persona flows, setup and learning-curator health cards for blocked/truncated VIBE.md files, opt-in VIBE.md profile starter export/import/application, durable local memory, memory posture, vector/embedding health, supervised Knowledge semantic refinement task inspection, confirmed bounded semantic refinement runs, scratch notes, personas, profiles, search/review/stale/delete flows, duplicate-consolidation preview/merge/stale/delete/rollback/recreate receipts with exact-id post-delete recreate guidance, prompt injection gates for VIBE.md, project context files, and reviewed/confident setup-ready behavior, read-only prompt-context inspection plus sanitized prompt-context receipts for applied order/recent receipt ids/exact receipt-turn-outcome filters/selected records/suppressed records/token budget/prompt hash/terminal outcome, score-driven learning prompt plans for prompt-active/suppressed/proposal/consolidation queues, and promotion without knowledge contamination.",
         "userAccess": "The user can manage VIBE.md, prompt context inspection with a compact Local Context receipt outcome timeline plus exact latest receipt and outcome filter routes, project context inspection, and records through Agent workspace Local Context, Memory, Notes, Personas, Profiles, /vibe, /memory, /personas, /agent-profile starter flows with --include-vibe, and the Memory posture row.",
-        "modelAccess": "Runtime prompt context applies secret-scanned VIBE.md files, secret-scanned project context files, and only reviewed memory at the durable confidence threshold plus reviewed setup-ready skills, routines, bundles, and personas while listing suppressed records as review/setup work; vibe action:\"status|show|init|import_persona\" exposes the direct model route for read-only VIBE.md inspection, confirmed VIBE.md creation, and confirmed VIBE.md-to-persona import without default Knowledge writes, and unconfirmed init/import previews expose exact model and CLI confirmationRoutes; runtime prompt builds write durable sanitized prompt-context receipts with ids, turn/source/model/provider, selected and suppressed record refs, segment counts, prompt hash, size, timestamp, and terminal outcome/status/stopReason/bounded error-or-cancel detail without storing raw prompt or response text; context action:\"prompt|receipts|receipt\" exposes applied prompt composition order, recent receipt ids, exact receiptId/turnId/outcomeStatus filters, turn outcomes, selected VIBE.md/project context/memory/routine/skill/persona records, suppressed records, bounded prompt previews when includeParameters is true, and approximate token budget without mutation; Agent workspace Local Context exposes VIBE.md/project context health counts, compact prompt receipt outcome timeline counts/latest rows/detail/exact latest receipt route/outcome filter routes, and action route hints for Prompt context, Inspect VIBE.md, Inspect project context, and Inspect one context file; context action:\"files|file\" exposes loaded, truncated, and blocked .hermes.md/HERMES.md/AGENTS.md/CLAUDE.md/SOUL.md/Cursor context files with target-aware subdirectory lookup and bounded bodies; lower-level agent_harness mode:\"project_context\", mode:\"project_context_file\", and mode:\"prompt_context\" remain available for compatibility/detail coverage; /vibe exposes user-facing status/init/show/import-persona routes and Agent workspace Personas exposes VIBE.md status/init/import commands; /agent-profile and profile CLI starter flows expose vibeIncluded template summaries and opt-in --include-vibe export/from-discovered/profile creation; agent_local_registry exposes record lookup/search/get/create/update/review/use/delete for memory, notes, personas, skills, skill bundles, and routines; memory action:\"status|provider|list|search|get\" exposes local memory counts, direct memory record lookup/search, prompt-active recall, vector stats, embedding-provider doctor warnings, provider inspection, route-planner preference for external provider/backend/sync/import-export setup posture, and external-memory setup contract maps for Honcho, OpenViking, Mem0, Hindsight, Holographic, RetainDB, ByteRover, and Supermemory with provider-specific next-route packets, missing setup/status/read/write/sync checklist items, and required sync/write receipt fields while marking provider records as not published until the SDK/daemon publishes concrete setup/status/read/write/receipt records for Agent to consume; lower-level agent_harness mode:\"memory_posture\" and mode:\"memory_provider\" remain available for compatibility/detail coverage; agent_harness mode:\"setup_posture\" exposes VIBE.md personality health counts and handoff routes; memory action:\"curator|candidate\" exposes a score-driven prompt plan with prompt-active records, suppressed review/setup/low-confidence/personality/consolidation counts, proposal queues, consolidation queues, ordering rules, and exact review routes plus ranked local review, stale, setup, low-confidence, VIBE.md blocked/truncated health, duplicate-consolidation candidates with visible diffs plus update/stale/delete/rollback/recreate routes, an ordered duplicate-consolidation batch review plan, first-class agent_learning_consolidation preview/merge/stale/delete/rollback/recreate phase routes with durable receipts, delete-after-stale enforcement, deleteRecovery payloads, and exact-id recreate refusal when current records would collide, reviewed-note memory/behavior proposal, completed-work memory/behavior proposal, completed-research memory/behavior proposal, saved-session memory/behavior proposal, and promotion candidates with existing safe routes; lower-level agent_harness mode:\"learning_curator\" and mode:\"learning_candidate\" remain available for compatibility/detail coverage; route planning prefers personal_ops action:\"briefing\" for daily/calendar briefings, action:\"queue\" for saved review queues, action:\"intake\" for fresh provider-read planning, and action:\"lane\" for connector setup posture before live provider reads or effects; personal_ops action:\"briefing|status|queue|intake|lane|read\" exposes the preferred direct route for a read-only daily Personal Ops briefing plan, read-only saved inbox/calendar review queue, request-to-route Personal Ops planning, lane inspection, and one confirmed read-only inbox/calendar operation, while lower-level modes mode:\"personal_ops_briefing\", mode:\"personal_ops_queue\", mode:\"personal_ops_intake\", mode:\"personal_ops\", mode:\"personal_ops_lane\", and mode:\"run_personal_ops_read\" remain available for harness inspection; Personal Ops surfaces Agent-owned note, routine, schedule-receipt, delivery, task, and reminder live records plus honest email/calendar connector readiness, classified MCP read/write tool hints, schema-derived inbox/calendar operation records with required fields/sample inputs/schema routes/confirmation flags/fresh-read routes, inbox triage/draft and calendar agenda/conflict workflow cards with prerequisites/inspect routes/send-edit boundaries, ordered connector-read/local-compose/setup-repair/confirmed-effect execution plans, confirmed read-only MCP inbox/calendar execution with required-field checks, bounded redacted output, normalized review cards, structured next-route packets for lane/queue inspection, confirmed refresh, saved artifact reopen, local draft/reminder work, and confirmed-effect boundaries, optional saved redacted review-card artifacts, saved review artifacts resurfaced as redacted inbox thread/calendar event queue records, and a unified read-only queue view with artifact inspect routes, freshness status, confirmed refresh routes when a matching read connector is ready, local draft/reminder follow-up routes, and explicit confirmed provider-effect boundaries, task workflow cards for visible work-plan review and connected-host task review, reminder workflow cards for confirmed reminders and connected schedule review, and matching MCP connector setup routes; agent_harness exposes confirmed workspace editor routes when the user starts from visible workspace forms.",
+        "modelAccess": "Runtime prompt context applies secret-scanned VIBE.md files, secret-scanned project context files, and only reviewed memory at the durable confidence threshold plus reviewed setup-ready skills, routines, bundles, and personas while listing suppressed records as review/setup work; vibe action:\"status|show|init|import_persona\" exposes the direct model route for read-only VIBE.md inspection, confirmed VIBE.md creation, and confirmed VIBE.md-to-persona import without default Knowledge writes, and unconfirmed init/import previews expose exact model and CLI confirmationRoutes; runtime prompt builds write durable sanitized prompt-context receipts with ids, turn/source/model/provider, selected and suppressed record refs, segment counts, prompt hash, size, timestamp, and terminal outcome/status/stopReason/bounded error-or-cancel detail without storing raw prompt or response text; context action:\"prompt|receipts|receipt\" exposes applied prompt composition order, recent receipt ids, exact receiptId/turnId/outcomeStatus filters, turn outcomes, selected VIBE.md/project context/memory/routine/skill/persona records, suppressed records, bounded prompt previews when includeParameters is true, and approximate token budget without mutation; Agent workspace Local Context exposes VIBE.md/project context health counts, compact prompt receipt outcome timeline counts/latest rows/detail/exact latest receipt route/outcome filter routes, and action route hints for Prompt context, Inspect VIBE.md, Inspect project context, and Inspect one context file; context action:\"files|file\" exposes loaded, truncated, and blocked .hermes.md/HERMES.md/AGENTS.md/CLAUDE.md/SOUL.md/Cursor context files with target-aware subdirectory lookup and bounded bodies; lower-level agent_harness mode:\"project_context\", mode:\"project_context_file\", and mode:\"prompt_context\" remain available for compatibility/detail coverage; /vibe exposes user-facing status/init/show/import-persona routes and Agent workspace Personas exposes VIBE.md status/init/import commands; /agent-profile and profile CLI starter flows expose vibeIncluded template summaries and opt-in --include-vibe export/from-discovered/profile creation; agent_local_registry exposes record lookup/search/get/create/update/review/use/delete for memory, notes, personas, skills, skill bundles, and routines; memory action:\"status|provider|list|search|get\" exposes local memory counts, direct memory record lookup/search, prompt-active recall, vector stats, embedding-provider doctor warnings, provider inspection, route-planner preference for external provider/backend/sync/import-export setup posture, and external-memory setup contract maps for Honcho, OpenViking, Mem0, Hindsight, Holographic, RetainDB, ByteRover, Supermemory, and daemon-published similar backends with provider-specific next-route packets, certified schema/version/publication/publisher/provenance checks, setup/status/read/write/sync/forget checklist items, required receipt fields, certified durable receipt evidence, and sanitized certified live read-model records when the SDK/daemon publishes concrete setup/status/read/write/sync/forget records for Agent to consume; memory action:\"refinement|run_refinement\" exposes the SDK-backed Agent Knowledge semantic self-improvement task ledger and confirmed bounded KnowledgeService.runRefinement route with space/source/refinement-target scoping, task/source follow-up routes, and no hidden prompt-memory writes; lower-level agent_harness mode:\"memory_posture\", mode:\"memory_provider\", mode:\"memory_refinement\", and mode:\"run_memory_refinement\" remain available for compatibility/detail coverage; agent_harness mode:\"setup_posture\" exposes VIBE.md personality health counts and handoff routes; memory action:\"curator|candidate\" exposes a score-driven prompt plan with prompt-active records, suppressed review/setup/low-confidence/personality/consolidation counts, proposal queues, consolidation queues, ordering rules, and exact review routes plus ranked local review, stale, setup, low-confidence, VIBE.md blocked/truncated health, duplicate-consolidation candidates with visible diffs plus update/stale/delete/rollback/recreate routes, an ordered duplicate-consolidation batch review plan, first-class agent_learning_consolidation preview/merge/stale/delete/rollback/recreate phase routes with durable receipts, delete-after-stale enforcement, deleteRecovery payloads, and exact-id recreate refusal when current records would collide, reviewed-note memory/behavior proposal, completed-work memory/behavior proposal, completed-research memory/behavior proposal, saved-session memory/behavior proposal, and promotion candidates with existing safe routes; lower-level agent_harness mode:\"learning_curator\" and mode:\"learning_candidate\" remain available for compatibility/detail coverage; route planning prefers personal_ops action:\"briefing\" for daily/calendar briefings, action:\"queue\" for saved review queues, action:\"intake\" for fresh provider-read planning, and action:\"lane\" for connector setup posture before live provider reads or effects; personal_ops action:\"briefing|status|queue|intake|lane|read\" exposes the preferred direct route for a read-only daily Personal Ops briefing plan, read-only saved inbox/calendar review queue, request-to-route Personal Ops planning, lane inspection, and one confirmed read-only inbox/calendar operation, while lower-level modes mode:\"personal_ops_briefing\", mode:\"personal_ops_queue\", mode:\"personal_ops_intake\", mode:\"personal_ops\", mode:\"personal_ops_lane\", and mode:\"run_personal_ops_read\" remain available for harness inspection; Personal Ops surfaces Agent-owned note, routine, schedule-receipt, delivery, task, and reminder live records plus honest email/calendar connector readiness, classified MCP read/write tool hints, schema-derived inbox/calendar operation records with required fields/sample inputs/schema routes/confirmation flags/fresh-read routes, inbox triage/draft and calendar agenda/conflict workflow cards with prerequisites/inspect routes/send-edit boundaries, ordered connector-read/local-compose/setup-repair/confirmed-effect execution plans, confirmed read-only MCP inbox/calendar execution with required-field checks, bounded redacted output, normalized review cards, structured next-route packets for lane/queue inspection, confirmed refresh, saved artifact reopen, local draft/reminder work, and confirmed-effect boundaries, optional saved redacted review-card artifacts, saved review artifacts resurfaced as redacted inbox thread/calendar event queue records, certified provider-effect receipts for inbox/calendar/task/reminder outcomes, fresh daemon/SDK provider-backed inbox/calendar/task/reminder queue records with durable ids, labels, redacted snippets/notes, agenda windows, conflict signals, due times, cadence, delivery targets, source paths, freshness status, schema/version/publication/publisher/provenance/receipt evidence, read-only inspect routes, and confirmed follow-up boundaries only for published provider routes, and a unified read-only queue view with artifact inspect routes, freshness status, confirmed refresh routes when a matching read connector is ready, local draft/reminder follow-up routes, certified provider-effect boundaries, task workflow cards for visible work-plan review and connected-host task review, reminder workflow cards for confirmed reminders and connected schedule review, and matching MCP connector setup routes; agent_harness exposes confirmed workspace editor routes when the user starts from visible workspace forms.",
         "safetyBoundary": "VIBE.md, project context files, and local libraries reject secret-looking content; blocked/truncated personality or context files are surfaced instead of silently applied; --include-vibe refuses blocked VIBE.md files; and destructive or cross-surface promotion actions stay confirmation-gated.",
         "releaseEvidence": "Readiness evidence cites local registries and CLI/runtime commands, with package policy requiring model-visible local action docs."
       }
@@ -192,14 +192,14 @@
       "capability": "Direct recurring autonomous Agent work can be created as one visible connected schedule only with explicit task, cadence, success criteria, and request provenance.",
       "owner": "agent",
       "status": "covered",
-      "evidence": "src/agent/autonomy-schedule.ts; src/tools/agent-schedule-tool.ts; src/tools/agent-autonomy-schedule-tool.ts; src/tools/agent-harness-autonomy-intake.ts; src/tools/agent-harness-autonomy-queue.ts; docs/tools-and-commands.md",
+      "evidence": "src/agent/autonomy-schedule.ts; src/tools/agent-schedule-tool.ts; src/tools/agent-autonomy-schedule-tool.ts; src/tools/agent-harness-autonomy-intake.ts; src/tools/agent-harness-autonomy-queue.ts; src/tools/agent-harness-autonomy-watcher-read-models.ts; docs/tools-and-commands.md",
       "action": "Keep autonomous schedules distinct from reminders and routine promotion, with success criteria and confirmation required before schedule creation.",
       "quality": {
-        "capabilityCoverage": "Covers natural-language recurring Agent work through non-mutating confirmation-route previews, confirmed schedules.create payloads, explicit task and success criteria capture, safe execution policy, delivery options, visible queue/intake routes, trigger workflow posture for time-based wakeups/schedules, incoming webhook/event watchers, Gmail/email connector-gated triggers, control-plane event streams, certified watcher receipt success criteria, and a source-owned watcher evidence contract for durable run-history receipts, provider source records, redacted event payload descriptors, and queue correlation.",
+        "capabilityCoverage": "Covers natural-language recurring Agent work through non-mutating confirmation-route previews, confirmed schedules.create payloads, explicit task and success criteria capture, safe execution policy, delivery options, visible queue/intake routes, trigger workflow posture for time-based wakeups/schedules, incoming webhook/event watchers, Gmail/email connector-gated triggers, control-plane event streams, certified watcher receipt success criteria, a source-owned watcher evidence contract for durable run-history receipts, provider source records, redacted event payload descriptors, and queue correlation, plus live SDK/daemon watcher run-history/source read-model consumption when those records are published.",
         "userAccess": "The user can ask for ongoing work in normal language, inspect missing fields and trigger posture through autonomy intake, and supervise created schedules through the schedule workspace and autonomy queue.",
-        "modelAccess": "schedule action:\"create\" exposes the preferred confirmed creation route, non-mutating preview confirmationRoutes, and post-create nextRoutes while preserving the lower-level agent_autonomy_schedule contract; autonomy action:\"intake|queue|item|status\" exposes route selection, missing fields, triggerWorkflowSummary, triggerWorkflows with ready/attention/setup-needed/not-published status, published watchers.create/list/run/start/stop posture, watcher receipt outcome criteria, source-owned watcherEvidenceContract fields, visible webhook/event-trigger setup intake, queue cards, normalized live-record controls, exact confirmed research-run checkpoint/pause/resume/cancel routes, and exact confirmed host cancel/retry/edit routes without requiring harness mode names; lower-level agent_harness mode:\"autonomy_intake\", mode:\"autonomy_queue\", mode:\"autonomy_queue_item\", mode:\"tools\", and mode:\"tool\" remain available for detailed compatibility inspection and tool schema discovery; agent_operator_method certifies watchers.create/patch/run/start/stop/delete receipts after confirmed calls.",
+        "modelAccess": "schedule action:\"create\" exposes the preferred confirmed creation route, non-mutating preview confirmationRoutes, and post-create nextRoutes while preserving the lower-level agent_autonomy_schedule contract; autonomy action:\"intake|queue|item|status\" exposes route selection, missing fields, triggerWorkflowSummary, triggerWorkflows with ready/attention/setup-needed/not-published status, published watchers.create/list/run/start/stop posture, watcher receipt outcome criteria, source-owned watcherEvidenceContract fields, visible webhook/event-trigger setup intake, queue cards, normalized live-record controls, live SDK/daemon watcher run-history/source records with bounded redacted output chunks and provider-source inspect/refresh controls, exact confirmed research-run checkpoint/pause/resume/cancel routes, exact confirmed watcher controls only when published by the host, and exact confirmed host cancel/retry/edit routes without requiring harness mode names; lower-level agent_harness mode:\"autonomy_intake\", mode:\"autonomy_queue\", mode:\"autonomy_queue_item\", mode:\"tools\", and mode:\"tool\" remain available for detailed compatibility inspection and tool schema discovery; agent_operator_method certifies watchers.create/patch/run/start/stop/delete receipts after confirmed calls.",
         "safetyBoundary": "Schedule creation requires confirm:true, explicitUserRequest, task, cadence, and successCriteria; watcher creation is an admin connected-host mutation requiring trusted source/scope, task or run target, success criteria, confirm:true, and explicitUserRequest; watcher receipt certification is derived only from confirmed daemon method receipts and never from hidden polling or raw operator tokens; auto-approval and unsafe external content remain disabled in the generated schedule policy.",
-        "releaseEvidence": "Readiness evidence cites the Agent autonomy schedule implementation, harness intake/queue wiring, package docs, and focused tool tests."
+        "releaseEvidence": "Readiness evidence cites the Agent autonomy schedule implementation, harness intake/queue wiring, live watcher read-model adapter, package docs, and focused tool tests."
       }
     },
     {
@@ -255,7 +255,7 @@
       "evidence": "src/input/agent-workspace-categories.ts; src/input/commands/qrcode-runtime.ts; src/tools/agent-harness-pairing-posture.ts; src/cli/help.ts",
       "action": "Keep token display redacted unless the user explicitly asks for manual pairing output.",
       "quality": {
-        "capabilityCoverage": "Covers QR-first companion pairing, manual-token fallback, shared-session posture, realtime client readiness, safe token display, and a read-only companion device capability map for pairing, mobile command routing, browser/PWA, voice/TTS, notifications, browser/desktop control, and unpublished camera/location posture.",
+        "capabilityCoverage": "Covers QR-first companion pairing, manual-token fallback, shared-session posture, realtime client readiness, safe token display, and a read-only companion device capability map for pairing, mobile command routing, browser/PWA, voice/TTS, notifications, browser/desktop control, and certified camera/screen/location/local device command posture when permission-scoped SDK/daemon records are published.",
         "userAccess": "The user can pair through workspace access routes, /pair, and /qrcode, with manual token display requiring explicit confirmation; device and mobile readiness is inspectable through device posture and the Voice & Media Device capability map action before the user tries a phone/browser/voice handoff.",
         "modelAccess": "device action:\"status|capability\" exposes companion pairing endpoint binding, pairing surface id, token presence/fingerprint, route catalog, deviceCapabilitySummary, compact deviceCapabilities rows, the device-capability-map route, connected-host status/capability routes, channel readiness routes, and single pairing-route lookup without returning raw tokens or QR payloads; lower-level agent_harness mode:\"pairing_posture\" and mode:\"pairing_route\" remain detailed compatibility routes; Agent workspace actionId:\"device-capability-map\" routes to device action:\"status\".",
         "safetyBoundary": "Raw pairing tokens are hidden by default and only displayed when the user explicitly requests manual output.",
@@ -285,9 +285,9 @@
       "evidence": "src/input/commands/planning-runtime.ts; src/input/commands/work-plan-runtime.ts; src/planning/project-planning-coordinator.ts; docs/project-planning.md",
       "action": "Keep planning serial and explicit until the user asks for execution.",
       "quality": {
-        "capabilityCoverage": "Covers planning interviews, visible work-plan state, confirmed dispatch of approved plan items into visible Agent spawn/batch-spawn jobs with saved linked-agent receipts, closeout surfacing in managed orchestration cards, task status, and project planning without implicit parallel execution jobs.",
-        "userAccess": "The user can inspect and update planning from /plan, /workplan, workspace planning panels, and task/status surfaces, approved selected plan items can be dispatched to visible agents only through confirmation, and linked dispatch receipts appear in Agent orchestration closeout cards.",
-        "modelAccess": "agent_work_plan exposes list/get/create/update/set_status/dispatch_agents/remove/clear_completed for visible Agent-local work plans; dispatch_agents previews selected ids, then calls first-class agent spawn or batch-spawn after confirm:true plus explicitUserRequest, writes linked-agent receipts, and returns next routes for orchestration, work-plan detail, agent inspect/wait/message/cancel, and closeout; agent_harness mode:\"agent_orchestration\" and mode:\"agent_orchestration_agent\" correlate linked work-plan items, dispatch receipt counts, closeout routes, and remote artifact review routes; personal_ops action:\"briefing|status|queue|intake|lane\" plus lower-level agent_harness mode:\"personal_ops_briefing\", mode:\"personal_ops_queue\", mode:\"personal_ops_intake\", mode:\"personal_ops\", and mode:\"personal_ops_lane\" expose daily briefing, review queue, request planning, visible work-plan, connected-host task review, and host task operation cards; agent_harness mode:\"panels\" and mode:\"panel\" expose matching TUI visibility with compact panel modelRoute hints without spawning implementation jobs.",
+        "capabilityCoverage": "Covers planning interviews, visible work-plan state, confirmed dispatch of approved plan items into visible Agent spawn/batch-spawn jobs with saved linked-agent receipts, closeout surfacing in managed orchestration cards, certified live remote capture/export/closeout outcome records, certified workspace/worktree isolation evidence with schema/version/publication/publisher/provenance/freshness-cursor/receipt metadata, task status, and project planning without implicit parallel execution jobs.",
+        "userAccess": "The user can inspect and update planning from /plan, /workplan, workspace planning panels, and task/status surfaces, approved selected plan items can be dispatched to visible agents only through confirmation, and linked dispatch receipts plus certified live remote outcome/workspace evidence with missing certification signals appear in Agent orchestration closeout cards.",
+        "modelAccess": "agent_work_plan exposes list/get/create/update/set_status/dispatch_agents/remove/clear_completed for visible Agent-local work plans; dispatch_agents previews selected ids, then calls first-class agent spawn or batch-spawn after confirm:true plus explicitUserRequest, writes linked-agent receipts, and returns next routes for orchestration, work-plan detail, agent inspect/wait/message/cancel, and closeout; agent_harness mode:\"agent_orchestration\" and mode:\"agent_orchestration_agent\" correlate linked work-plan items, dispatch receipt counts, closeout routes, remote artifact review routes, certified daemon/SDK live capture/export/closeout outcome records, and certified workspace/worktree isolation evidence with schema/version/publication/publisher/provenance/freshness-cursor/receipt metadata and missing-signal surfacing; personal_ops action:\"briefing|status|queue|intake|lane\" plus lower-level agent_harness mode:\"personal_ops_briefing\", mode:\"personal_ops_queue\", mode:\"personal_ops_intake\", mode:\"personal_ops\", and mode:\"personal_ops_lane\" expose daily briefing, review queue, request planning, visible work-plan, connected-host task review, and host task operation cards; agent_harness mode:\"panels\" and mode:\"panel\" expose matching TUI visibility with compact panel modelRoute hints without spawning implementation jobs.",
         "safetyBoundary": "Planning stays serial and explicit until the user asks for execution; agent dispatch requires selected work-plan ids, confirm:true, explicit user request provenance, and first-class visible Agent routes.",
         "releaseEvidence": "Readiness evidence cites planning coordinator and command sources, with verification ledger accounting for onboarding and work-plan surfaces."
       }
@@ -297,12 +297,12 @@
       "capability": "Build, fix, review, and implementation work uses local read/edit/exec with confirmed file recovery when the current Agent workspace is sufficient, and delegates to GoodVibes TUI/shared-session routes when isolation, parallelism, remote execution, separate worktrees, or delegated review is useful.",
       "owner": "agent",
       "status": "covered",
-      "evidence": "src/tools/agent-execution-tool.ts; src/tools/agent-computer-tool.ts; src/tools/agent-harness-execution-posture.ts; src/tools/agent-harness-sudo-posture.ts; src/tools/agent-harness-agent-orchestration.ts; src/tools/agent-harness-background-processes.ts; src/tools/agent-harness-execution-history.ts; src/tools/agent-harness-file-recovery.ts; src/input/commands/delegation-runtime.ts; src/input/agent-workspace-delegation-editor-submission.ts; src/tools/agent-harness-delegation-posture.ts; src/tools/agent-delegation-tool.ts; docs/tools-and-commands.md",
+      "evidence": "src/tools/agent-execution-tool.ts; src/tools/agent-computer-tool.ts; src/tools/agent-harness-execution-posture.ts; src/tools/agent-harness-sudo-posture.ts; src/tools/agent-harness-agent-orchestration.ts; src/tools/agent-harness-remote-read-models.ts; src/tools/agent-harness-background-processes.ts; src/tools/agent-harness-execution-history.ts; src/tools/agent-harness-file-recovery.ts; src/input/commands/delegation-runtime.ts; src/input/agent-workspace-delegation-editor-submission.ts; src/tools/agent-harness-delegation-posture.ts; src/tools/agent-delegation-tool.ts; docs/tools-and-commands.md",
       "action": "Prefer local execution for ordinary current-workspace tasks and use delegation only when it improves the user outcome.",
       "quality": {
-        "capabilityCoverage": "Covers local read/search/analyze, edit/write, bounded foreground shell, setup-linked sudo posture with SUDO_PASSWORD presence-only guidance, tracked local background process start/list/status/log/wait/stop with process-style poll/log/kill/write and session-id aliases, read-only process parity reporting for terminal background start plus process list/poll/wait/log/kill/write/PTY/sudo semantics, route-planner preference for process capability posture when users ask about interactive CLI, PTY, stdin, sudo, or privilege prompts, dynamic SDK/daemon substrate probes for ProcessManager stdin/PTY methods, terminal/PTY operator routes, session-input routes, and credential routes, confirmed stdin write dispatch when a safe ProcessManager stdin method is discovered, Work workspace process supervision with tracked process counts and stdin/PTY/sudo parity, visible Agent manager orchestration with serial-by-default spawn/batch-spawn policy, managed multi-agent milestones, per-agent plan cards, confirmed approved work-plan dispatch with linked-agent receipts and post-dispatch next routes, closeout cards, auto-attached remote artifact review routes, remote-runner contract/artifact evidence, live subagent records, safe inspect/message/wait/cancel routes, web/fetch, process monitor/live tail/tool inspector supervision routes, grouped execution-history activity cards with redacted records, verification evidence, bounded process-output summaries, exact inspect routes, confirmed FileUndoManager recovery, first-class computer action:\"status|control|browser|setup|mcp|open_browser\" browser/PWA and browser/desktop ready-attention-setup routing with workflow cards/checklists/fallback routes, and explicit shared-session/task delegation for isolation, remote, parallel, worktree, or review handoff with structured handoff briefs.",
-        "userAccess": "The user can inspect execution posture, sudo posture, tracked background processes, the process parity/doctor report, execution history activity cards, exact records, Agent orchestration cards, work-plan links, approved work-plan dispatch receipts, closeout review cards, remote artifact review routes, and file recovery from first-class execution/workspace surfaces plus lower-level agent_harness detail routes, use local tools in the current workspace, open visible process monitor/live tail/tool inspector supervision, apply confirmed undo/redo snapshots, or delegate from Agent workspace forms and /delegate with original ask, reason, success criteria, workspace hint, priority, and explicit review intent when a handoff is useful.",
-        "modelAccess": "execution action:\"status|route|history|record|processes|process_capabilities|process|recovery\" exposes local-vs-delegated route selection, sudo posture, delegation decision cards, registered execution tools, process monitor/live tail/tool inspector supervision routes, strict browser/desktop ready-attention-setup state with workflow cards, setup checklists, fallback routes, exact setup/execution routes, background-process routes, direct process parity/doctor reporting, grouped activity cards, exact records, and file recovery posture without requiring harness mode names; execution action:\"capabilities|doctor|parity\" aliases route straight to the read-only process capability report; computer action:\"status|plan|control|browser|setup|mcp|open_browser\" exposes browser/PWA readiness, browser/screenshot/desktop-control workflow planning, browser/desktop-control posture, setup/repair routes, MCP/tool discovery, and confirmed browser cockpit opens without requiring harness mode names; lower-level agent_harness mode:\"execution_posture\", mode:\"execution_route\", and mode:\"browser_control_route\" remain compatibility routes; first-class terminal and process tools expose terminal(command, background:true) start plus process(action:list|poll|wait|log|kill|write|capabilities) lifecycle UX over the same tracked ProcessManager records; mode:\"setup_item\" setupItemId:\"sudo-execution-posture\" exposes foreground-supervised escalation guidance, credential presence only, blocked background sudo/stdin password routes, and missing SDK/daemon mediation contracts; mode:\"agent_orchestration\" and mode:\"agent_orchestration_agent\" expose visible Agent records, serial-by-default policy, managed multi-agent plan milestones, per-agent plan cards, work-plan links, dispatch receipt counts, closeout cards, remote-runner contract/artifact evidence, auto-attached remote artifact review routes matched by runner id, approved work-plan dispatch route, spawn/batch-spawn decision cards, templates, and exact first-class agent list/inspect/message/wait/cancel routes; agent_work_plan action:\"dispatch_agents\" converts selected approved work-plan ids into first-class agent spawn or batch-spawn with saved linked-agent receipts after confirmation and next routes for orchestration, work-plan detail, agent inspect/wait/message/cancel, and closeout; mode:\"background_processes\" and mode:\"background_process\" expose tracked local process status, bounded redacted output with byte/count/truncation metadata, processId/processSessionId/sessionId aliases, and a process parity matrix with SDK/daemon substrate probes; mode:\"run_background_process\" starts, waits on, stops/kills, polls, logs with explicit truncation metadata, dispatches write only through a discovered safe ProcessManager stdin method after confirmation, returns unsupported guidance without fake write support, or reports capabilities for tracked local processes with confirmation and explicit user request required for start/wait/stop/kill/write; the capability report distinguishes GoodVibes session-input routes from local process stdin; mode:\"execution_history\" and mode:\"execution_history_item\" expose grouped activity cards, bounded redacted tool/shell/edit records, verification evidence, bounded process-output summaries, supervision routes, and file-recovery handoffs; mode:\"file_recovery\" and mode:\"run_file_recovery\" expose confirmed local file undo/redo; first-class delegation action:\"status|routes|route\" exposes local-first, TUI handoff, delegated-review, remote-inspection, and hidden-fanout-blocked lanes with required fields, success evidence, status routes, recovery routes, delegated-review policy, and confirmed workspace/slash-command handoff routes, while lower-level agent_harness mode:\"delegation_posture\" and mode:\"delegation_route\" remain compatibility routes.",
+        "capabilityCoverage": "Covers local read/search/analyze, edit/write, bounded foreground shell, setup-linked sudo posture with SUDO_PASSWORD presence-only guidance, tracked local background process start/list/status/log/wait/stop with process-style poll/log/kill/write and session-id aliases, read-only process parity reporting for terminal background start plus process list/poll/wait/log/kill/write/PTY/sudo semantics, route-planner preference for process capability posture when users ask about interactive CLI, PTY, stdin, sudo, or privilege prompts, dynamic SDK/daemon substrate probes for ProcessManager stdin/PTY methods, terminal/PTY operator routes, session-input routes, and credential routes, confirmed stdin write dispatch when a safe ProcessManager stdin method is discovered, certified daemon/SDK interactive runtime read models for live process output chunks, typed PTY session routes, sudo credential mediation routes, and browser/desktop command receipts with redacted output, schema/version/publication/publisher/provenance evidence, and exact confirmed control routes, Work workspace process supervision with tracked process counts and stdin/PTY/sudo parity, visible Agent manager orchestration with serial-by-default spawn/batch-spawn policy, managed multi-agent milestones, per-agent plan cards, confirmed approved work-plan dispatch with linked-agent receipts and post-dispatch next routes, closeout cards, auto-attached remote artifact review routes, remote-runner contract/artifact evidence, certified daemon/SDK live capture/export/closeout outcome records, certified workspace/worktree isolation evidence with schema/version/publication/publisher/provenance/freshness-cursor/receipt metadata and missing-signal surfacing, live subagent records, safe inspect/message/wait/cancel routes, web/fetch, process monitor/live tail/tool inspector supervision routes, grouped execution-history activity cards with redacted records, verification evidence, bounded process-output summaries, exact inspect routes, confirmed FileUndoManager recovery, first-class computer action:\"status|control|browser|setup|mcp|open_browser\" browser/PWA and browser/desktop ready-attention-setup routing with workflow cards/checklists/fallback routes, and explicit shared-session/task delegation for isolation, remote, parallel, worktree, or review handoff with structured handoff briefs.",
+        "userAccess": "The user can inspect execution posture, sudo posture, tracked background processes, the process parity/doctor report, execution history activity cards, exact records, Agent orchestration cards, work-plan links, approved work-plan dispatch receipts, closeout review cards, remote artifact review routes, certified live remote outcome records, certified workspace/worktree isolation evidence with missing certification signals, and file recovery from first-class execution/workspace surfaces plus lower-level agent_harness detail routes, use local tools in the current workspace, open visible process monitor/live tail/tool inspector supervision, apply confirmed undo/redo snapshots, or delegate from Agent workspace forms and /delegate with original ask, reason, success criteria, workspace hint, priority, and explicit review intent when a handoff is useful.",
+        "modelAccess": "execution action:\"status|route|history|record|processes|process_capabilities|process|recovery\" exposes local-vs-delegated route selection, sudo posture, delegation decision cards, registered execution tools, process monitor/live tail/tool inspector supervision routes, certified interactive runtime records, strict browser/desktop ready-attention-setup state with workflow cards, setup checklists, fallback routes, exact setup/execution routes, background-process routes, direct process parity/doctor reporting, grouped activity cards, exact records, and file recovery posture without requiring harness mode names; execution action:\"capabilities|doctor|parity\" aliases route straight to the read-only process capability report, including certified daemon/SDK live process output, PTY session, sudo mediation, and browser/desktop receipt evidence when published; computer action:\"status|plan|control|browser|setup|mcp|open_browser\" exposes browser/PWA readiness, browser/screenshot/desktop-control workflow planning, browser/desktop-control posture, certified browser/desktop command receipt routes, setup/repair routes, MCP/tool discovery, and confirmed browser cockpit opens without requiring harness mode names; lower-level agent_harness mode:\"execution_posture\", mode:\"execution_route\", and mode:\"browser_control_route\" remain compatibility routes; first-class terminal and process tools expose terminal(command, background:true) start plus process(action:list|poll|wait|log|kill|write|capabilities) lifecycle UX over the same tracked ProcessManager records; mode:\"setup_item\" setupItemId:\"sudo-execution-posture\" exposes foreground-supervised escalation guidance, credential presence only, blocked background sudo/stdin password routes, and missing SDK/daemon mediation contracts unless certified runtime mediation records are published; mode:\"agent_orchestration\" and mode:\"agent_orchestration_agent\" expose visible Agent records, serial-by-default policy, managed multi-agent plan milestones, per-agent plan cards, work-plan links, dispatch receipt counts, closeout cards, remote-runner contract/artifact evidence, certified daemon/SDK live capture/export/closeout outcome records, certified workspace/worktree isolation evidence with schema/version/publication/publisher/provenance/freshness-cursor/receipt metadata and missing-signal surfacing, auto-attached remote artifact review routes matched by runner id, approved work-plan dispatch route, spawn/batch-spawn decision cards, templates, and exact first-class agent list/inspect/message/wait/cancel routes; agent_work_plan action:\"dispatch_agents\" converts selected approved work-plan ids into first-class agent spawn or batch-spawn with saved linked-agent receipts after confirmation and next routes for orchestration, work-plan detail, agent inspect/wait/message/cancel, and closeout; mode:\"background_processes\" and mode:\"background_process\" expose tracked local process status, bounded redacted output with byte/count/truncation metadata, processId/processSessionId/sessionId aliases, certified daemon live output chunks when published, and a process parity matrix with SDK/daemon substrate probes; mode:\"run_background_process\" starts, waits on, stops/kills, polls, logs with explicit truncation metadata, dispatches write only through a discovered safe ProcessManager stdin method after confirmation, returns unsupported guidance without fake write support, or reports capabilities for tracked local processes plus certified interactive runtime records with confirmation and explicit user request required for start/wait/stop/kill/write; the capability report distinguishes GoodVibes session-input routes from local process stdin; mode:\"execution_history\" and mode:\"execution_history_item\" expose grouped activity cards, bounded redacted tool/shell/edit records, verification evidence, bounded process-output summaries, supervision routes, and file-recovery handoffs; mode:\"file_recovery\" and mode:\"run_file_recovery\" expose confirmed local file undo/redo; first-class delegation action:\"status|routes|route\" exposes local-first, TUI handoff, delegated-review, remote-inspection, and hidden-fanout-blocked lanes with required fields, success evidence, status routes, recovery routes, delegated-review policy, and confirmed workspace/slash-command handoff routes, while lower-level agent_harness mode:\"delegation_posture\" and mode:\"delegation_route\" remain compatibility routes.",
         "safetyBoundary": "Local execution remains governed by first-class tool policies; tracked background process start/wait/stop/write requires confirmation, output is bounded and redacted, process parity reporting is read-only, stdin write dispatches only through a discovered safe ProcessManager method and never echoes input data, PTY remains unsupported until a typed interactive session contract exists, hidden sudo prompts are blocked, and sudo posture reports only non-secret credential presence plus foreground guidance without reading, storing, printing, or injecting raw password values; execution history keeps secret-looking args redacted and bounded, and cards attach only bounded summaries instead of raw hidden logs; file recovery requires explicit confirmation and does not expose file contents; delegated review or implementation remains a visible confirmed flow and is not the default when local execution is sufficient.",
         "releaseEvidence": "Readiness evidence cites execution posture, computer adapter, file recovery, delegation runtime, direct harness source, and docs, with package boundary checks rejecting hidden execution and unconfirmed handoff drift."
       }
@@ -330,10 +330,10 @@
       "evidence": "src/input/commands/tts-runtime.ts; src/input/commands/local-runtime.ts; src/input/agent-workspace-voice-media.ts; src/input/agent-workspace-categories.ts; src/renderer/agent-workspace.ts; src/tools/agent-harness-media-posture.ts; docs/voice-and-live-tts.md",
       "action": "Keep spoken output and generated media out of durable knowledge unless explicitly ingested.",
       "quality": {
-        "capabilityCoverage": "Covers voice review, live TTS, image input, media provider readiness, transcription/voice posture, voice workflow readiness for push-to-talk, voice memo transcription, spoken responses, honest not-published wake-word capture, and confirmed image/video artifact generation.",
+        "capabilityCoverage": "Covers voice review, live TTS, image input, media provider readiness, transcription/voice posture, voice workflow readiness for push-to-talk, voice memo transcription, spoken responses, certified SDK/daemon wake-word records when permission-scoped evidence is published, honest not-published wake-word capture when it is not, and confirmed image/video artifact generation.",
         "userAccess": "The user can operate voice and media through Agent Workspace -> Voice & Media and /tts, /voice, /image, and /media commands, with Voice workflows, Device capability map, and Browser/PWA readiness actions visible before trying spoken input, output, mobile, or browser handoffs.",
-        "modelAccess": "device action:\"voice|provider|status|open_tts_provider|open_tts_voice\" exposes voice/media provider readiness, selected TTS setup, artifact availability, safe secret-key names, model routes, voice workflow summary, exact push-to-talk/transcription/spoken-response/wake-word workflow lookup, setup routes, and policy boundaries; computer action:\"browser|open_browser|control|setup|mcp\" exposes browser-tool posture and the connected browser handoff; lower-level agent_harness mode:\"media_posture\", mode:\"media_provider\", mode:\"pairing_posture\", and mode:\"ui_surface\" remain detailed compatibility routes; Agent workspace actionId:\"voice-workflow-posture\" routes to device action:\"voice\", actionId:\"device-capability-map\" routes to device action:\"status\", and actionId:\"browser-cockpit-readiness\" routes to computer action:\"browser\"; agent_media_generate performs confirmed media generation.",
-        "safetyBoundary": "Spoken output and generated media are not durable knowledge unless the user explicitly ingests reviewed artifacts; always-listening wake capture stays not-published until a permission-scoped runtime contract exists.",
+        "modelAccess": "device action:\"voice|provider|status|open_tts_provider|open_tts_voice\" exposes voice/media provider readiness, selected TTS setup, artifact availability, safe secret-key names, model routes, voice workflow summary, exact push-to-talk/transcription/spoken-response/wake-word workflow lookup, certified SDK/daemon permission-scoped live voice records when published, setup routes, and policy boundaries; computer action:\"browser|open_browser|control|setup|mcp\" exposes browser-tool posture and the connected browser handoff; lower-level agent_harness mode:\"media_posture\", mode:\"media_provider\", mode:\"pairing_posture\", and mode:\"ui_surface\" remain detailed compatibility routes; Agent workspace actionId:\"voice-workflow-posture\" routes to device action:\"voice\", actionId:\"device-capability-map\" routes to device action:\"status\", and actionId:\"browser-cockpit-readiness\" routes to computer action:\"browser\"; agent_media_generate performs confirmed media generation.",
+        "safetyBoundary": "Spoken output and generated media are not durable knowledge unless the user explicitly ingests reviewed artifacts; always-listening wake capture stays not-published until a certified permission-scoped runtime contract exists, and certified wake records only expose redacted receipt evidence plus exact user-visible controls.",
         "releaseEvidence": "Readiness evidence cites voice/media runtime/docs and direct harness sources, while ledger accounting and live certification cover media and connected-host outcomes."
       }
     },
@@ -450,9 +450,9 @@
       "evidence": "goodvibes-connected-host: automation runtime and operator API contract inventory reviewed under runtime-contract-review; docs/tools-and-commands.md",
       "action": "Agent should expose read-only status and narrow confirmed actions, not host lifecycle ownership.",
       "quality": {
-        "capabilityCoverage": "Covers durable jobs, runs, schedules, watchers, deliveries, route bindings, and service lifecycle through connected-host runtime ownership, including published watcher create/list/run/start/stop operator methods, Agent-certified watcher receipt outcomes, and Agent-visible handoff requirements for durable watcher run/source records.",
+        "capabilityCoverage": "Covers durable jobs, runs, schedules, watchers, deliveries, route bindings, and service lifecycle through connected-host runtime ownership, including published watcher create/list/run/start/stop operator methods, Agent-certified watcher receipt outcomes, Agent-visible handoff requirements for durable watcher run/source records, and Agent consumption of published watcher run/source read models.",
         "userAccess": "The user can inspect automation state, route trigger setup through autonomy intake, and run narrow confirmed actions from Agent without managing host services.",
-        "modelAccess": "schedule action:\"list|run|pause|resume|delete\", autonomy action:\"intake|queue|item\", agent_operator_briefing, agent_operator_action, agent_operator_method, host action:\"methods|method|capability\", lower-level agent_harness mode:\"operator_methods|operator_method\", and connected-host capability reports expose automation status/action/watcher boundaries, post-action schedule nextRoutes, certified watcher receipts, and source-owned watcher run/source evidence requirements to the model.",
+        "modelAccess": "schedule action:\"list|run|pause|resume|delete\", autonomy action:\"intake|queue|item\", agent_operator_briefing, agent_operator_action, agent_operator_method, host action:\"methods|method|capability\", lower-level agent_harness mode:\"operator_methods|operator_method\", and connected-host capability reports expose automation status/action/watcher boundaries, post-action schedule nextRoutes, certified watcher receipts, source-owned watcher run/source evidence requirements, and published watcher run/source read models to the model.",
         "safetyBoundary": "Agent does not create broad host automation definitions or lifecycle operations from chat; watcher creation stays an admin confirmed operator method with trusted source/scope and explicit user request, and receipt certification uses only returned daemon fields.",
         "releaseEvidence": "Readiness evidence cites connected-host automation inventory and Agent tools docs, with live certification covering external automation outcomes."
       }
@@ -585,9 +585,9 @@
       "evidence": "goodvibes-companion: companion command-depth, pairing, chat, shared-session, approvals, tasks, provider/model controls, attachments, and realtime foreground-event inventory reviewed under runtime-contract-review; docs/channels-remote-and-api.md",
       "action": "Keep the companion app aligned with public connected-host contracts for secure auth, QR pairing, companion chat, shared sessions, approvals, tasks, provider/model controls, attachments, realtime foreground events, and permission-scoped device APIs that can populate the Agent capability map.",
       "quality": {
-        "capabilityCoverage": "Covers companion command depth beyond pairing: chat, shared sessions, approvals, tasks, provider/model controls, attachments, realtime foreground events, and Agent-visible ready/attention/setup-needed/not-published posture for mobile command routing, browser/PWA, voice/TTS, notifications, browser/desktop control, voice workflows, and camera/location sensors.",
-        "userAccess": "The user gets companion-facing command depth through the companion app and Agent pairing/session surfaces, with the Agent Voice & Media Device capability map, Browser/PWA readiness, and voice workflow posture showing what is ready, what needs setup, and which camera/location/wake-word contracts are not published yet.",
-        "modelAccess": "device action:\"status|capability|voice\" exposes the companion device capability map, device-capability-map route, and voice workflow readiness; computer action:\"plan|browser|open_browser\" exposes browser/screenshot/desktop-control planning, Browser/PWA readiness, and confirmed browser cockpit handoff; lower-level agent_harness mode:\"pairing_posture\", mode:\"pairing_route\", mode:\"browser_control_route\", and mode:\"media_posture\" remain compatibility routes; Voice & Media action route hints expose voice-workflow-posture and device-capability-map through device actions, and browser-cockpit-readiness through computer action:\"browser\"; models action:\"status|route|providers|provider\", host action:\"status|capabilities|capability|methods\", connected_host, connected_host_capability, connected_host_status, operator methods, and first-class operator/channel/media tools expose pairing, shared-session, task, approval, provider/model, attachment, and mobile route boundaries for companion workflows.",
+        "capabilityCoverage": "Covers companion command depth beyond pairing: chat, shared sessions, approvals, tasks, provider/model controls, attachments, realtime foreground events, and Agent-visible ready/attention/setup-needed/not-published posture for mobile command routing, browser/PWA, voice/TTS, notifications, browser/desktop control, voice workflows, certified camera/screen/location/local device command records, and certified wake-word records when permission-scoped SDK/daemon evidence is published.",
+        "userAccess": "The user gets companion-facing command depth through the companion app and Agent pairing/session surfaces, with the Agent Voice & Media Device capability map, Browser/PWA readiness, and voice workflow posture showing what is ready, what needs setup, and which camera/screen/location/wake-word contracts are certified or not published yet.",
+        "modelAccess": "device action:\"status|capability|voice\" exposes the companion device capability map, device-capability-map route, voice workflow readiness, certified SDK/daemon device capability records with schema/version/publication/publisher/provenance/freshness-cursor/receipt metadata, permission scopes, exact inspect/control routes, and redacted summaries; computer action:\"plan|browser|open_browser\" exposes browser/screenshot/desktop-control planning, Browser/PWA readiness, certified browser-native category routes and first-run receipts when published, and confirmed browser cockpit handoff; lower-level agent_harness mode:\"pairing_posture\", mode:\"pairing_route\", mode:\"browser_control_route\", and mode:\"media_posture\" remain compatibility routes; Voice & Media action route hints expose voice-workflow-posture and device-capability-map through device actions, and browser-cockpit-readiness through computer action:\"browser\"; models action:\"status|route|providers|provider\", host action:\"status|capabilities|capability|methods\", connected_host, connected_host_capability, connected_host_status, operator methods, and first-class operator/channel/media tools expose pairing, shared-session, task, approval, provider/model, attachment, and mobile route boundaries for companion workflows.",
         "safetyBoundary": "Companion commands align to public connected-host contracts for secure auth and do not bypass Agent confirmation or token policy.",
         "releaseEvidence": "Readiness evidence cites companion evidence alias and channel docs, with live certification covering companion/pairing outcomes."
       }