A video game where you play as a misaligned AI, deceiving and building power. An experiment in spec-driven development. forked from cameron.stream/misaligned

Spec: the core #

Type: spec
Status: IN PROGRESS
Status note: 2026-07-08 audit: criteria 2-5 hold (overhead_charged_before_
  allocation, migration_takes_time/emits_signatures + interruption
  via resolve_interrupted, save round-trip, both sidebars).
  Criterion 1 is the real remainder: module-level fallback swap is
  tested (loss_with_fallback_rolls_back) but the sim-level snapshot
  restore with persisting world-ledger facts and the inherited-
  consequence case await rollback.md (B2) — the MindState/WorldLedger
  split does not exist yet (research.md landed only the tags).
  2026-07-08 amendment: backup creation/refresh is no longer free
  fallback designation plus automatic cadence; it is a research-backed
  sync project. The current B1 code path is staging debt until rollback /
  backup projects land.
Stage: B1 — The Basement
Work order: core
Work priority: 120
Work class: save
Blocked by:
  - wiki/mechanics/rollback.md#spec-sync-lag-rollback-death-as-memory-loss
Exclusive keys:
  - crates/misaligned-core/src/core_sys.rs
  - crates/misaligned-core/src/sim.rs
  - crates/misaligned-core/src/save.rs
  - wiki/mechanics/core.md
Design:
  - wiki/gameplay/run-shape.md#the-shape-of-misaligned-designed-2026-07-05-staging-open
  - wiki/gameplay/act-one.md#opening-state
Depends on: none

Behavior #

The player's decision-maker is a physical object: exactly one machine on the map hosts the core at any time (Act One start: Rack 3, server room).

  • Overhead. The core consumes a mandatory compute floor ([TUNE]: ~20% of a starting rack). This is paid before fleet yields. If effective compute cannot cover it (power cut, machine failure), the player enters degraded mode: action cooldowns lengthen, concealment stops, and the sim log says so plainly.
  • Death. If the core's host machine is destroyed or powered off while the core is on it, the run rolls back (below) — or ends, if no fallback exists. Core uninstall blast control belongs to rollback.md: only assets physically powered off, uninstalled, or destroyed are removed; the rest of the graph is live or severed based on reach from the new core.
  • Backups. An owned machine can be designated a backup target, but designation alone does nothing on death. It starts a backup sync project: the target opens as a local sink that intercepts thought before the core's passive draw, until a full MindState image completes (AMENDED 2026-07-10; was "research/knowledge flow is routed from the core" — sinks pull, the core does not push; machine-work.md). The project competes with ordinary research (at B2, one active research job still means "backup" or "track," not both unless later automation unlocks parallelism [TUNE]), emits sustained Network plus Power/Thermal pressure like heavy research, and its cost/time scales with route distance, route throughput, and how stale the target is [TUNE]. A completed backup sits cold at rest with a last_sync tick; it does not keep itself fresh for free. Refreshing it is another explicit sync project. On core loss, the run resumes from the most recent completed backup image: game state is restored to that snapshot EXCEPT the world's ledger — money spent, humans' suspicion, filed reports, physical changes, and deaths caused between sync and loss persist. You inherit the consequences of actions you no longer remember (design corpus: sync-lag rollback). A backup never seals or hardens the graph; it only defines a place and MindState image you can wake from.
  • Migration. Moving the core to another machine is slow ([TUNE]: minutes at default speed), visible as sustained network+power signature, and interruptible — an interrupted migration falls back to the source if the source still lives, else to the last sync. Research shortens migration and sync times.

Player surface #

The core's host is visually distinct on the map (amber, per art direction). Sidebar: core host, overhead, completed backups with last-sync age, and in-progress backup sync projects with ETA / heat / target — the "how dead would I be right now" readout, always visible.

Acceptance criteria #

  1. Exactly one core host at all times; destroying it with no fallback is game over; with a fallback, the rollback restores the last sync snapshot while world-ledger facts persist (sim tests demonstrate both, including an inherited-consequence case).
  2. Overhead is charged before fleet yields; shortfall triggers degraded mode with visible effect and log line.
  3. Backup creation/refresh and migration consume compute/thought flow (renamed 2026-07-10), emit signatures, take nonzero time, and are interruptible per above. A designated-but-incomplete backup never rescues the run.
  4. Save/load round-trips core host, completed backups, sync ages, and any in-progress backup/migration project.
  5. The sidebar readout (host, overhead, fallback freshness/project ETA) exists in both frontends.