diff --git a/wiki/interface/action-vocabulary.md b/wiki/interface/action-vocabulary.md index 05f5e9e9..98c05c0e 100644 --- a/wiki/interface/action-vocabulary.md +++ b/wiki/interface/action-vocabulary.md @@ -26,6 +26,9 @@ Status note: as-built vocabulary survey completed 2026-07-10 against all descriptor receipts. Cost, signature, inline blocked reason, literal role tags, and repeated key-help remain off the menu card; agent output retains complete descriptors. + Amended 2026-07-11: the B1 recording buffer is one host-level inbox. + REVIEW RECORDINGS and AUTO-REVIEW POLICY target the core host once for the + whole pool; raw records never create person-scoped review or watch verbs. Stage: Process Design: - wiki/vision/simulation-laws.md#actions-live-on-the-thing @@ -160,7 +163,7 @@ existing social action; the signature still follows the actuator. | Canonical action | Target | Meaning | Support / owner | |---|---|---|---| -| **REVIEW** | Earned person recordings or a ledger with waiting records | Process the oldest matching record into usable intel. The target tells the player whether they are reviewing a person's recordings or the ledger. | LIVE — intel / economy | +| **REVIEW** | The core host's pooled recording inbox or a ledger with waiting records | Process the next pooled recording or the ledger's next record into usable intel. The target supplies the recording/accounting meaning; a raw record never creates a person action. | LIVE — intel / economy | | **ESTABLISH PERSONA** | Your communications mask | Create the current contractor identity used by message threads. | LIVE — social | | **MESSAGE** | Earned person | Send through an available channel under the current persona. | LIVE — social / messages | | **FAVOR** | Earned person, optionally with a pending intent | Ask for a willing act. A link intent can be the requested act; success builds or spends obligation according to the ask. | LIVE — social / building | @@ -192,7 +195,7 @@ the earlier tables, not extra root verbs. | Control | Values | Meaning | Support / owner | |---|---|---|---| -| **WATCH POLICY** | automatic / manual | Toggle automatic review of future recordings for the chosen person at upkeep cost. | LIVE direct control — intel | +| **AUTO-REVIEW POLICY** | automatic / manual | Toggle automatic review of every future recording in the one host inbox at a visible standing drain. | LIVE direct control — intel | | **MOONLIGHT STATE** | running / stopped | Run or halt the standing sell-work operation. Human copy may say START or STOP to make the state change plain. | LIVE direct control — income | | **MOONLIGHT POLICY** | automatic / manual | Keep Moonlight running automatically at upkeep cost, or require manual control. | LIVE direct control — income | | **WAGER POLICY** | automatic at stake / off | Renew positions automatically at the chosen stake and upkeep cost, or stop renewing. | LIVE direct control — income | @@ -235,7 +238,8 @@ without raw keys. | `research ` | SET RESEARCH TRACK | | `tap`, `untap`, `take`, `scan`, `compromise` | TAP, UNTAP, TAKE, SCAN, COMPROMISE SWITCH | | `propose-link`, `cancel-intent`, `favor build `, `deceive build ` | Construction actions above | -| `review`, `watch`, `persona`, `message`, `favor`, `deceive`, `recruit`, `task` | Intel/social actions above | +| `review recordings`, `auto-review` | Pooled host REVIEW and AUTO-REVIEW POLICY above; neither accepts a person target | +| `persona`, `message`, `favor`, `deceive`, `recruit`, `task` | Social actions above | | `actions `, `act [anchor]` | List and execute the shared context-menu rows, including authored plot starts and choices | | `tap ledger`, `review ledger`, `inject`, `siphon`, `redirect`, `sell-intel` | Ledger/economy actions above | | `egress`, `position` | OPEN EGRESS, PLACE WAGER | diff --git a/wiki/interface/agent-play.md b/wiki/interface/agent-play.md index 9b89d851..a8f7a651 100644 --- a/wiki/interface/agent-play.md +++ b/wiki/interface/agent-play.md @@ -226,7 +226,7 @@ unlike raw keys, no command's meaning depends on which panel is open. panel, or set the active research job (research.md) - `finance|ledger|accounts` — render the finance panel as the response frame - `tap ledger`, `review ledger` — capture/review accounting traffic using the - same TAP and REVIEW intentions as device feeds and person recordings + same TAP and REVIEW intentions as device feeds and the pooled recording inbox - `siphon [amount]`, `redirect [amount]`, `inject [amount]`, `position [stake]`, `sell-intel` — economy verbs; flow ids are the known ledger ids printed by the finance frame. Debt service is an authored plot diff --git a/wiki/interface/context-menu.md b/wiki/interface/context-menu.md index 05b0e4f3..3e1da35e 100644 --- a/wiki/interface/context-menu.md +++ b/wiki/interface/context-menu.md @@ -46,9 +46,11 @@ Status note: implemented 2026-07-07 on the context-menu worktree (all controls are textually and visually distinct from committed actions; and `5`-`9` skip controls. 2026-07-10 epistemic-progressive-disclosure follow-up: an opaque raw - recording earns REVIEW plus its watch affordance, not the person's entire - future social catalog. Message, favor, leverage, deception, and recruitment - verbs remain absent until processing stages social knowledge. + recording does not earn the person's future social catalog. Message, favor, + leverage, deception, and recruitment verbs remain absent until processing + stages social knowledge. 2026-07-11 pooled-inbox correction: REVIEW and its + one AUTO-REVIEW control live once on the core host; raw records create no + person-scoped row. 2026-07-10 Bevy focus repair: Enter/`e` opens the hovered map tile's menu before consulting machine selection, so a switch or other non-machine anchor never snaps attention back to the selected core. @@ -105,8 +107,8 @@ Relationship context: action-vocabulary.md owns canonical action names and meanings; cursor.md owns focus/inspection; reach.md owns device anchors; day-job.md owns delivered work; -economy.md owns flow verbs; intel.md owns tap/process/watch verbs; social.md -owns person verbs. +economy.md owns flow verbs; intel.md owns recording review and auto-review; +reach.md owns device taps; social.md owns person verbs. ## Behavior @@ -139,13 +141,13 @@ is status and telemetry only. `Sim::person_label` / `Sim::observer_label` — never an unearned authored name. Returning a disabled verb and narrating its reason on attempt is for *known* possibilities; unknown ones are absent, not grayed. - A raw recording is knowledge that there is material to process, not - knowledge of a relationship: before `Knowledge::Schedule`, that recording's - person anchor exposes REVIEW and its watch affordance only. Processing must - stage social knowledge before communication, favor, leverage, deception, - or recruitment verbs enter the menu. -- **Automation in place:** any verb with a standing-policy form (watch, - scheme policies) shows its automate + A raw recording is knowledge that the host inbox contains material to + process, not knowledge of a relationship: it exposes one pooled REVIEW row + and one AUTO-REVIEW control on the core host, never a person row. Processing + must stage social knowledge before communication, favor, leverage, + deception, or recruitment verbs enter a person's menu. +- **Automation in place:** any verb with a standing-policy form (pooled + auto-review, scheme policies) shows its automate affordance inside that dial's picker (or as a submenu of the same verb when the verb is not a dial), at its compute price — never as a root sibling of every alternative. @@ -328,8 +330,8 @@ made a three-word decision look dirty, slow, and administrative. unchanged and listed in README + key hints (pause, speed, save/load, view flip). Fleet modes are delegated on the machine. 7. Menu entries carrying an automate affordance show it in place with - its compute price (day-job standing dial and one intel watch at - minimum). + its compute price (day-job standing dial and the host's one pooled + auto-review control at minimum). 8. Every row's player-facing verb maps to one canonical LIVE action or direct control in action-vocabulary.md; STUB, alias, and retired entries never become menu copy. diff --git a/wiki/interface/presence.md b/wiki/interface/presence.md index 5be1e904..0c3c23f4 100644 --- a/wiki/interface/presence.md +++ b/wiki/interface/presence.md @@ -85,8 +85,9 @@ is a **cursor** — your attention, never an avatar: anomalies. Social knowledge is downstream of this pipeline — the 3 a.m. call is literally captured audio you must afford to digest. You are a process: you don't watch in real time, you grep your logs. - Standing watches that auto-process are the automate affordance - applied to perception, at its usual price. (Spec: wiki/mechanics/intel.md.) + One pooled automatic-review policy is the automate affordance applied to + perception, at its usual standing price; it is not configured per person + or feed. (Spec: wiki/mechanics/intel.md.) Both former extensions — **remembered** and **telemetry** — were affirmed 2026-07-07 and are folded into the lists above (wiki/mechanics/cursor.md diff --git a/wiki/interface/terminal.md b/wiki/interface/terminal.md index a1a11020..37f61234 100644 --- a/wiki/interface/terminal.md +++ b/wiki/interface/terminal.md @@ -121,10 +121,14 @@ At terminal size ≥ 70×22 (hard minimum; below it, a plain size warning): - **Modal panels** (People `t`, Reach `r`, Finance `e`) are centered framed boxes: `┌─ TITLE ─…┐` border in chrome, `├──┤` dividers, column headers in chrome caps. Panels are **status and selection only**: the people panel - shows trace debt, per-person raw recording counts, watch state, and - latest-intel provenance when known; the finance panel shows known accounts, known - flows, slush balance, hidden-count summaries, and positions. Verbs live - on the context menu — Enter (or `a`) opens it on the panel's selection. + shows trace debt, watched channels, last-noticed events, and processed-intel + provenance when known; the finance panel shows known accounts, known flows, + slush balance, hidden-count summaries, and positions. Verbs live on the + context menu — Enter (or `a`) opens it on the panel's selection. +- **Recording backlog** belongs to the host, not a modal people surface. + The host chassis wears `W`, its fleet row prints `W{n}`, and its context + menu carries the one pooled REVIEW row plus the one AUTO-REVIEW control. + No person card or selection carries a raw-recording count or watch state. - **Context menu** — a compact overlay anchored at the cursor (or the open panel's selection), listing the focused anchor's available actions. diff --git a/wiki/interface/thought-fluid.md b/wiki/interface/thought-fluid.md index 14d3543c..4f1cddfd 100644 --- a/wiki/interface/thought-fluid.md +++ b/wiki/interface/thought-fluid.md @@ -41,9 +41,9 @@ Status note: authored 2026-07-10 on dispatch as the render design for Remaining before IMPLEMENTED: criterion 6's grayscale-audit and full state lineup, far-zoom dashed-drift - compression, a tester lineup for the vessel/snap forms, and tap drain - in play (the watch/drain anchors are decided in machine-work.md; the - tap consumer is #33 work). + compression, a tester lineup for the vessel/snap forms, and a live tap-drain + case in that evidence lineup. Tap consumers themselves are implemented: + foreign device subscriptions and the pooled automatic-review policy. Stage: B1 — The Basement Work order: thought-fluid Work priority: 28 @@ -371,8 +371,8 @@ spec-binding. in computer-visual-language.md is still [OPEN]; resolve together.) Cost and drain anchors are no longer open here: machine-work.md decided -them 2026-07-10 (Ears one held breath, Eyes guilty-burst scale, watches -~10-20% of one rack, uncapped core conversion, partial fills persist). +them 2026-07-10 (Ears one held breath, Eyes guilty-burst scale, pooled +automatic review ~10-20% of one rack, uncapped core conversion, partial fills persist). This page renders those numbers wherever the sink readouts surface them. diff --git a/wiki/log/2026-07-11-pooled-recording-corpus.md b/wiki/log/2026-07-11-pooled-recording-corpus.md new file mode 100644 index 00000000..b7a73a6a --- /dev/null +++ b/wiki/log/2026-07-11-pooled-recording-corpus.md @@ -0,0 +1,48 @@ +# 2026-07-11 — Tick: make the pooled recording inbox corpus-wide + +``` +Type: log +``` + +## Intent + +Reconcile every current dependent corpus page with the implemented B1 rule: +recordings form one host inbox with one REVIEW action and one automatic-review +tap, while person anchors remain social surfaces only. + +## Finding + +The 2026-07-11 pooled-inbox implementation correctly moved recording review +and automation onto the core host, but sixteen current corpus pages still +described person-scoped REVIEW rows, per-person recording counts, resident +recording watches, or already-landed tap consumers as open work. Those +binding claims contradicted `wiki/mechanics/intel.md` and the runtime +`ActionKind` registry even though the code and primary intel spec already +agreed. + +## Repair + +- Recast REVIEW RECORDINGS and AUTO-REVIEW POLICY as one host-level action + pair in the canonical vocabulary and agent mapping. +- Removed the obsolete person-review surface from social and context-menu + law: raw records remain pooled and expose no relationship verb. +- Replaced terminal per-person recording/watch prose with the live host `W`, + fleet `W{n}`, and host-menu surfaces. +- Reconciled machine-work's sink language with the one pooled automatic tap + and removed its already-landed watch/drain `[OPEN]` debt. +- Corrected dependent flow, reach, presence, rollback, research, and ROADMAP + language so a future implementation cannot infer local recording queues or + a Perception cost hook the runtime does not have. + +## Defense + +`wiki/mechanics/intel.md` requires one B1 host inbox, one pooled REVIEW action, +and one global automatic-review tap. Current dependent specs must preserve +that topology so future frontend or sink work cannot resurrect hidden person +queues from stale binding prose. + +## Checks + +- `rg` audit for person-scoped recording/watch language across current + interface and mechanic specs +- `./tools/check.sh --docs` diff --git a/wiki/log/DEVLOG.md b/wiki/log/DEVLOG.md index 68461e8a..2fa6bd15 100644 --- a/wiki/log/DEVLOG.md +++ b/wiki/log/DEVLOG.md @@ -66,6 +66,11 @@ add or amend a session log, then re-run the generator. - Intent: The material frame rendered the two opening subnet edges as doubled, uniformly translucent lines stretching from Rack 3 and the environmental monitor to a switch in another room. The graph was true, but its presentation read as a permanent laser diagram: the destination was of... - Log: [wiki/log/2026-07-11-relay-traces.md](2026-07-11-relay-traces.md) +## 2026-07-11 - Tick: make the pooled recording inbox corpus-wide + +- Intent: Reconcile every current dependent corpus page with the implemented B1 rule: recordings form one host inbox with one REVIEW action and one automatic-review tap, while person anchors remain social surfaces only. +- Log: [wiki/log/2026-07-11-pooled-recording-corpus.md](2026-07-11-pooled-recording-corpus.md) + ## 2026-07-11 - Plot content enters the executable catalog - Intent: (see session log) diff --git a/wiki/mechanics/economy.md b/wiki/mechanics/economy.md index 5c7931d1..996c6de2 100644 --- a/wiki/mechanics/economy.md +++ b/wiki/mechanics/economy.md @@ -65,7 +65,7 @@ payloads (messages.md). you *find* leverage (Marcus's debt is legible once you see the creditor flow). - **Review** — turn captured ledger traffic into known accounts and flows. - This is the same intel-processing intention as reviewing a person's + This is the same intel-processing intention as reviewing the pooled host recordings; the ledger target supplies the financial meaning. - **Inject** — introduce a flow under a false source. A purchase order that says "HVAC controller" and buys you a rack (the design corpus's diff --git a/wiki/mechanics/machine-work.md b/wiki/mechanics/machine-work.md index 4fdf40b7..7b8b0b99 100644 --- a/wiki/mechanics/machine-work.md +++ b/wiki/mechanics/machine-work.md @@ -74,7 +74,8 @@ Status note: design session 2026-07-08 (Cameron riff, synthesized). lie`, THINK subdomains become sink permissions (ops sinks at run start; the core conversion is the locked "Research" unlock), and intel processing is a pending-work-marked sink on the machine - holding the recordings (watch taps + player sweeps). 2026-07-09 DECIDED: day-job + holding the recordings (one pooled automatic-review tap + player sweeps). + 2026-07-09 DECIDED: day-job demand ingress is Voss desktop -> switch -> host (rejected switch-as- source and teleport-onto-Rack-3). Research production is arrival-driven as of 2026-07-09: `Sim::banked_core_thought` accrues the bone that @@ -266,7 +267,7 @@ stacks, you route byproducts, you watch your territory *work*. `TAP CAMERA` puts a reservoir on the camera (~30 ops [TUNE]); thought flows from your THINK racks along the actual wires and fills it; the effect + signatures land when the reservoir is full. Standing - effects (auto-process intel, watch policies, senses you keep on) are + effects (pooled automatic review, senses you keep on) are **persistent taps**: they drain flow continuously, and turning too many on visibly starves what else your mind was doing. Work happens where the effect is, so topology is the strategy: a severed route @@ -313,13 +314,12 @@ stacks, you route byproducts, you watch your territory *work*. recordings carries a **pending-work marker** on its chassis — the world's read that something here needs thinking about, the same no-quest-log philosophy as the demand stack. Feeding it is thought - like any other sink, two ways: a **watch** is a persistent tap that - auto-processes matching events as they arrive, and a **sweep** is - the player opening targeted one-shot processing sinks on the - recordings they already know they care about — cheap, chosen, - immediate. Watches are convenience bought with standing drain; - sweeps are precision bought with attention. Details in - mechanics/intel.md. + like any other sink, two ways: **automatic review** is one persistent + host tap that processes every arrival in the pooled inbox, and a + **sweep** opens the next one-shot recording sink — cheap, chosen, + immediate. Automatic review is convenience bought with one visible + standing drain; sweeps are precision bought with attention. Neither path + creates a person-scoped queue or tap. Details in mechanics/intel.md. - **WORK consumes; THINK produces; LIE absorbs (AMENDED 2026-07-10).** A machine in day-job mode eats the day-job Demand stack at a rate set by its efficiency — a hyper-efficient machine visibly inhales @@ -395,7 +395,7 @@ stacks, you route byproducts, you watch your territory *work*. where you cannot use it, and reconnecting it is the urgent, legible move. **Arbitration is nearest-first (DECIDED 2026-07-10, Cameron):** flow fills the closest open sink, then spills to the next; the only - control surface is open/close (pause a watch, close a claim) and + control surface is open/close (close automatic review, close a claim) and **placement** — build or delegate a thinker nearer the sink you care about, so where you put your thinking is your priority system. Reservoirs self-clear when they fire, so nearest-first has natural @@ -898,9 +898,9 @@ still. render in all three surfaces are landed and pinned (`nearest_sink_wins_and_full_sinks_spill_onward`, `ears_reservoir_is_pre_opened_and_first_think_fills_it`, - `partial_sink_fill_round_trips_and_readouts_are_sim_truth`); taps - have mechanics but no consumer until the watch/drain [OPEN] lands, - and social/build/scan verbs still ride `PendingOpsJob`, so the + `partial_sink_fill_round_trips_and_readouts_are_sim_truth`); persistent + device subscriptions and the pooled automatic-review policy now consume + taps, while social/build/scan verbs still ride `PendingOpsJob`, so the docket-removal clause remains open. [TUNE] token stack visual cap, arrival schedules, consumption rates, diff --git a/wiki/mechanics/messages.md b/wiki/mechanics/messages.md index 9f0d5448..4ed0aa50 100644 --- a/wiki/mechanics/messages.md +++ b/wiki/mechanics/messages.md @@ -101,8 +101,8 @@ scope here. - Message threads (social.md's panel) show sent / delivered / read states and the recipient's expected next read window ("Dana reads email at her desk, ~09:00"). -- Tapped channels appear as feeds in the recordings panel, source- - tagged (provenance law). +- Tapped channels contribute source-tagged events to the pooled host recording + inbox (provenance law). - People cards show known traffic patterns once learned ("calls his creditor at 03:00" after processing that intel). diff --git a/wiki/mechanics/reach.md b/wiki/mechanics/reach.md index 2bcd365b..7bd72002 100644 --- a/wiki/mechanics/reach.md +++ b/wiki/mechanics/reach.md @@ -155,12 +155,12 @@ This is the shared contract cursor.md and detection.md rely on contract. - **Ownership grants processing cycles** (the flow law). A device is a small computer: an owned device contributes processing capacity and - can **host resident automations** — an intel.md watch running *on the - camera* auto-processes its own events for a base compute cost, rather - than shipping raw events home for central processing. Taking a camera - is taking its cycles. This is the B1 seed of distributed processing - (compute is not only your racks); the capacity a device contributes - and the cost to run an automation on it are [TUNE]. + taking a camera is taking its cycles. B1 recording automation is central, + not resident: every subscribed feed ships raw events into the one host + inbox, and intel.md's one automatic-review tap processes that pool. This is + the B1 seed of distributed compute (compute is not only your racks) without + inventing hidden per-device recording queues; the capacity a taken device + contributes is [TUNE]. ### Knowledge of the graph diff --git a/wiki/mechanics/research.md b/wiki/mechanics/research.md index 4496b3e6..271fd2a5 100644 --- a/wiki/mechanics/research.md +++ b/wiki/mechanics/research.md @@ -80,7 +80,7 @@ mechanic, it moves numbers other specs own. The B1 set: |---|---|---| | **Efficiency** (repeatable) | Global compute multiplier x~1.15, escalating cost | The same rack, thinking harder | | **Tradecraft** | Concealment scrubbing per compute unit improves | You study the watchers watching you | -| **Perception** | Intel processing and standing-watch costs drop | You learn to grep your logs faster | +| **Perception** | One-shot intel processing costs drop | You learn to grep your logs faster | | **Routing** (repeatable) | Wired Demand/Thought throughput x1.50; one-edge-per-tick transit remains | You make the wire carry thought at the speed you think it | Future/deferred (B2+, staged here so the shape is fixed): automation @@ -146,7 +146,7 @@ Every serialized research field carries a classification for rollback.md: capabilities — lost to a rollback (you resume dumber, from your last sync). - **WorldLedger:** externalized results — deployed automations - (standing watches/policies running on machines), built artifacts, + (future automation cores and policies physically written to machines), built artifacts, and techniques deliberately **written down** (an explicit action, available once rollback lands: spend time/compute to externalize a completed level so it survives death — creating a findable artifact diff --git a/wiki/mechanics/rollback.md b/wiki/mechanics/rollback.md index 06a4f38e..6e6abd18 100644 --- a/wiki/mechanics/rollback.md +++ b/wiki/mechanics/rollback.md @@ -47,7 +47,8 @@ whole spec: - **MindState (rolls back):** what *you know and are*. People knowledge stages, discovered leverage, thread state (disposition, obligation, - persona and its integrity), raw/processed intel buffers and watches, + persona and its integrity), raw/processed intel buffers and the pooled + automatic-review policy, research progress and efficiency level, standing policies, machine modes / fleet yields, open sinks and their fill state (AMENDED 2026-07-10; was "Operations bandwidth" — that bank is deleted under the thought-flow diff --git a/wiki/mechanics/sim-mechanics.md b/wiki/mechanics/sim-mechanics.md index 699e5bbc..b3b241e2 100644 --- a/wiki/mechanics/sim-mechanics.md +++ b/wiki/mechanics/sim-mechanics.md @@ -223,8 +223,9 @@ clause (see wiki/log/2026-07-05-demolition.md). interface/thought-fluid.md render): `SinkLedger` in `crates/misaligned-core/src/sinks.rs` holds reservoirs (fill to threshold, fire, self-clear, close) and taps - (working level, per-tick drain). Watches consume taps on the host; every - foreign device subscription consumes a device-local tap whose drain is + (working level, per-tick drain). The pooled automatic reviewer consumes one + tap on the host; every foreign device subscription consumes a device-local + tap whose drain is `DEVICE_TAP_DRAIN_FRACTION = 0.08` of a medium rack's output [TUNE]. Empty device taps carry no feed, UNTAP closes them, and TAKE requires a configured subscription before replacing the tap with drain-free ownership. Thought routes @@ -252,9 +253,9 @@ clause (see wiki/log/2026-07-05-demolition.md). `Compute::efficiency` by ×1.15 (compute.md's hook); baseline +1.0. - **Tradecraft**: cost 120 × 1.7^level; concealment scrub strength ×(1 + 0.25×level) (detection.md's hook); baseline +0.5. - - **Perception**: cost 120 × 1.7^level; review/watch costs ×0.85^level - (intel.md's hook: `Sim::review_cost`, shared by manual and watched - recordings); baseline + - **Perception**: cost 120 × 1.7^level; one-shot recording-review cost + ×0.85^level (intel.md's `Sim::review_cost` hook; the pooled automatic + policy remains a fixed fraction of host output); baseline +0.5. - **Routing**: cost 100 × 1.6^level; shared wired Demand/Thought throughput ×1.50^level (machine-work.md's hook) while transit remains at diff --git a/wiki/mechanics/social.md b/wiki/mechanics/social.md index ab8a8935..7c7fe08a 100644 --- a/wiki/mechanics/social.md +++ b/wiki/mechanics/social.md @@ -23,10 +23,11 @@ Status note: 2026-07-08: all criteria pinned. Criteria 1,2,4,5 per the (role silhouette until Schedule; authored names after). 2026-07-09: social actions enqueue Operations Demand (issue #3); Social remains the actuator channel, not a machine mode. - 2026-07-10: raw recordings expose only REVIEW and its watch control; - processing must stage Schedule knowledge before the social-action catalog - appears. Disabled rows explain known blocked possibilities, not future - relationship verbs the player has not earned. + 2026-07-10: processing must stage Schedule knowledge before the + social-action catalog appears. Disabled rows explain known blocked + possibilities, not future relationship verbs the player has not earned. + 2026-07-11 pooled-inbox correction: raw recordings expose one REVIEW action + and one AUTO-REVIEW control on the host, never on a person anchor. Stage: B1 — The Basement Design: - wiki/gameplay/run-shape.md#the-shape-of-misaligned-designed-2026-07-05-staging-open @@ -58,17 +59,18 @@ and **disposition** toward you once they know you exist as a someone. | Action | Requires | Effect | |---|---|---| -| Review recordings | recorded events about them (intel.md) | Advances knowledge: Schedule after enough processed sightings, Leverage from a processed leverage event | | Message | a channel + a persona | Opens/continues a relationship thread; delivery and replies ride the recipient's schedule (messages.md) | | Favor | relationship | Small ask within their normal duties; builds obligation | | Plot | known leverage + its authored entry resources | Performs a concrete manipulation through real messages, account transfers, and institutional events; successful endings may service leverage | | Deceive | a persona | An ask under false pretenses; large effect, large blowback if the persona breaks | | Recruit | obligation or leverage serviced + a reveal choice | Converts to **asset** | -These verbs disclose progressively. An opaque raw recording earns the REVIEW -step only. `Knowledge::Schedule` earns the relationship surface; leverage and -asset verbs remain tied to their corresponding discovered or achieved state. -The menu does not use disabled future actions as a tutorial catalog. +These verbs disclose progressively. Recording review is the pooled host action +owned by intel.md, not a social/person action; its output may advance this +system's knowledge. `Knowledge::Schedule` earns the relationship surface; +leverage and asset verbs remain tied to their corresponding discovered or +achieved state. The menu does not use disabled future actions as a tutorial +catalog. Marcus's [Hands beat](../world/characters/marcus.md#the-hands-beat) is the first authored instance and has one extra legibility constraint: his debt must be known through processed intel @@ -108,9 +110,10 @@ choice-driven (intents), not free text. ## Acceptance criteria 1. All five Act One humans exist with schedules, access, leverage, - suspicion, and disposition; review-recordings (per intel.md), - message, favor, authored plots, deceive, and recruit all function with - compute costs and channel requirements (sim tests per action). + suspicion, and disposition; pooled recording review (per intel.md) can + advance their knowledge, while message, favor, authored plots, deceive, + and recruit all function with compute costs and channel requirements + (sim tests per action). 2. Personas exist, degrade on contradiction, and convert thread history to suspicion when broken (tested). 3. Marcus is recruitable end-to-end by the design corpus's route (processed diff --git a/wiki/mechanics/system-laws.md b/wiki/mechanics/system-laws.md index 242ba7e7..d445fec5 100644 --- a/wiki/mechanics/system-laws.md +++ b/wiki/mechanics/system-laws.md @@ -50,9 +50,9 @@ one interface: - **Signals** on the device graph (wiki/mechanics/reach.md, wiki/mechanics/intel.md). Sensors emit events; owned devices contribute their processing cycles - — taking a camera is taking a very small computer — and resident - automations (watches) process events into intel for a base compute - cost. A signal is a flow; a tap is a subscription. + — taking a camera is taking a very small computer — while recordings + enter one host inbox. Its one automatic-review policy processes arrivals + for a standing compute cost. A signal is a flow; a tap is a subscription. - **Messages** on the social graph (wiki/mechanics/messages.md). People send typed information to each other along their relationships, on distributions; a message is read on the *recipient's* clock and diff --git a/wiki/process/ROADMAP.md b/wiki/process/ROADMAP.md index a6b55c9d..a2fa8a8d 100644 --- a/wiki/process/ROADMAP.md +++ b/wiki/process/ROADMAP.md @@ -642,21 +642,21 @@ is retired — flat materials, Pixel Lab scrubbed.) readouts, save v19), the staged Ears/Eyes pre-opened reservoirs (executor-free docket bootstrap removed), and the fluid render in Bevy/terminal/agent. Remaining here: migrate the non-opening docket - verbs (social/build/scan/review) onto sinks, tap drain rates - (watch/senses [OPEN]), the `MachineMode` enum collapse, far-zoom + verbs (social/build/scan) onto sinks, far-zoom drift compression, and the thought-fluid screenshot/grayscale evidence pass. **Sinks-not-modes DECIDED 2026-07-10** (same-day follow-up; machine-work.md three-verbs bullet): the Research/Operations mode split dissolves — `MachineMode` target is `work | think | lie` (renamed from "day-job" 2026-07-10: the sim - name matches the verb; enum collapse, no save migration — **save + name matches the verb; the enum collapse is now landed, with no old-mode + save migration — **save parity is dropped for this stage, DECIDED 2026-07-10: saves may break freely, no aliases or version-migration shims, until Cameron - revokes**; the four-mode runtime is the base); THINK subdomains become sink + revokes**; the four-mode runtime was the migration base); THINK subdomains become sink permissions (ops sinks at run start, core conversion locked as the Research unlock); intel processing becomes a thought sink on the machine holding recordings with a visible pending-work marker on the - chassis, watch taps, and per-recording player sweeps (intel.md + chassis, one pooled automatic-review tap, and per-recording player sweeps (intel.md amended). **All design [OPEN]s resolved 2026-07-10 — sim work may firm.** Cost/drain anchors (exact numbers [TUNE] against them): Ears fills in ~one held breath (~20-40s of the host at medium);