Sprint 4 Design Pack — GWP-283

Executive Summary

Out of 15 candidate concepts, 9 are keepers (strong KN-86 fit, distinct from the 17 launch modules, viable as cartridges in their current shape), 3 are conditional keepers (need rework before they’re sprintable), and 3 are cuts (overlap, scope, or “great fiction, wrong device”). The single most important read across the set: the strongest keepers all share a Lisp-native data shape — nested trees, sets-with-membership, lambdas-as-NPCs. The weakest entries are the ones that would feel native on any handheld and don’t lean on the device’s structural-editing paradigm. Use that as the filter for any future ideation pass.

The five-cart “first expansion wave” if we ever sprint cart development beyond launch:

WETWARE — clearest “you couldn’t make this on Switch” identity.
SIGNALWAKE — instant ‘88 recognition; doubles as marketing surface area.
RESURRECTION — the most KN-86-shaped narrative experience.
GHOSTLINE — Lisp-native canon piece; the ad-friendly “you walked through a memory tree” demo.
DAEMON — only cart in the set that runs between sessions; opens new gameplay design space the launch library doesn’t touch.

Headlines for Josh:

GARGOYLE and BLACK SUN are cuts — both fight the device’s solo, terminal-shaped premise.
HOSTILE TAKEOVER and RECITATION are conditional keepers — strong concepts but need a second pass on what the player does each minute.
TACHIKOMA is a quiet sleeper — agents-as-lambdas you edit across sessions is the most paradigm-faithful pitch in the bunch and worth a rev for the record.

Concepts that survived the filter (9 keepers)

1. WETWARE — Back-Alley Ripperdoc

Pitch: Graft cybernetic implants onto a patient’s anatomical tree. Power budget, immune response, neural compatibility — every implant rebalances the whole tree.
Mechanic hook: the patient is the data structure. You’re a structural editor on a body. Reroute a nerve to free a slot for a stronger optic; the gameplay verb is (splice ...).
KN-86 fit: highest — this is what the device’s tree-editing grammar was made to portray. The CIPHER voice carries the patient’s complaints in real time; that’s the secondary loop.
Phase/economy fit: Clean fit. Patient walks in (mission template), you operate (phase 1 — diagnose, phase 2 — install, phase 3 — close), reputation in implant scenes increases or tanks. Credit flow obvious. Multi-cart contracts: a CLONESHOP forger can require a WETWARE biometric anchor.
Risk/scope: Anatomical-tree authoring is real content — needs ~30 implant types, ~15 patient archetypes, ~6 complications. Medium-large cart by content volume.

2. SIGNALWAKE — Pirate BBS Sysop

Pitch: Run a pirate BBS in 1988. Forums, door games, ANSI gallery, war-dialer defense, contraband zines.
Mechanic hook: the cart is a BBS. You log in to your own board through the deck and curate it; users post, fights break out in threads, the FCC sniffs around. Time advances continuously while the cart is loaded.
KN-86 fit: highest by aesthetic — pure terminal experience, period-correct, 100% recognizable to anyone over 35. This is also the cart the marketing trailer leads with.
Phase/economy fit: Reputation = uptime + curation quality + contraband gravity. Credit flow = sysop tips + zine sales. Phase chain: nightly cycle (users post overnight, you moderate at “morning”). Long, ambient — slot orthogonal to mission carts.
Risk/scope: Content-heavy in a different way — needs a deep procedural-text generator for forum posts, ANSI art, door-game results. Lean on existing CIPHER procedural grammar.

3. RESURRECTION — Digital Necromancer

Pitch: Reconstruct a dead person’s personality from their digital exhaust. The client is grieving. You’re not bringing them back — you’re building the best lie you can.
Mechanic hook: assemble a Lisp lambda (the personality) from fragments of evidence (emails, voice clips, transaction logs). You’re literally editing a structural representation of a soul. Every choice you make echoes in the CIPHER voice playback. Sessions end when the client either accepts the reconstruction or breaks down.
KN-86 fit: most KN-86-shaped narrative experience — quiet, intimate, terminal-bound, deeply Lisp-shaped, emotionally heavy in a way only this device pulls off.
Phase/economy fit: Reputation in emotional-truth contracts (no other cart in the launch library has this affinity). Credits flow modestly — this is a prestige cart, not a money cart. Phase chain: gather → assemble → present.
Risk/scope: Heavy writing. Needs at least 6 client archetypes with hand-authored backstories so the procedural reassembly has real material to draw on. This is the most expensive cart in the keepers set per content-author hour. Worth it for the prestige tier.

4. GHOSTLINE — Cyber-forensic autopsy (GitS)

Pitch: Walk a compromised cyberbrain’s memory tree looking for the intruder’s footprint. Each memory is a node; tampering shows as pattern noise; trace the chain back to entry.
Mechanic hook: depth-first search through a player-explorable tree where the interesting moves are subtree comparison and pattern-matching against known intrusion signatures. The Lisp REPL is the in-fiction tool the operator uses.
KN-86 fit: inherently nested-list activity — exactly the shape the device’s grammar was built for. Ad-friendly: shows off the structural editor as gameplay.
Phase/economy fit: Forensics-affinity reputation (new). Credits per case + bonus on full-trace. Phase: scan → diff → attribute. Multi-phase missions with WETWARE (forensics on a botched implant) or BLACK LEDGER (forensics on a financial intrusion) write themselves.
Risk/scope: Content-light per case (procedural intrusion patterns generate from a tight grammar) but cases need believable memory content. ~15 case archetypes, scenario data. Medium cart.

5. DAEMON — Lisp sprites you write (Snow Crash)

Pitch: Author Lisp lambdas that run between play sessions — overnight, while the device is off. Wake up to results: contracts scanned, prices watched, targets surveilled.
Mechanic hook: the only cart in the set whose loop spans sessions. You write a daemon, you sleep, you check the results. Failure modes are great: a daemon mis-written with bad sentinel conditions burns its budget on garbage.
KN-86 fit: single most paradigm-faithful pitch — Lisp as the literal verb of play. Also: it’s the cart that proves the device’s premise (the deck thinks for you when you’re away).
Phase/economy fit: Daemons cost ENERGY (a new resource budget) per overnight tick; ENERGY refills per real-day. Reputation per successful daemon outcome. Multi-cart: a SIGNALWAKE moderator daemon, a BLACK LEDGER market-watch daemon — every other cart can host daemon-type sub-missions.
Risk/scope: Hardest to playtest because the loop spans real time. Need an emulator-side time-warp for QA. Worth the effort for the design-space expansion. Engineering risk: persistence model and daemon execution lifecycle aren’t free.

6. NIGHTOWL — Aerial Drone Pilot

Pitch: Pilot a drone over the Sprawl. Cityscape as a nested location tree; battery is the timer; the world is what you can see from above.
Mechanic hook: location-tree traversal under a hard timer. Visual: ASCII top-down city map. Mission types: surveillance, package drop, evasion.
KN-86 fit: Good but not highest. The location tree is grammar-shaped; the real-time battery timer is the tension. The terminal-as-cockpit framing carries it.
Phase/economy fit: Direct-action reputation (similar to ICE Breaker’s heat). Credits per delivery / surveillance hit. Phase: brief → fly → return.
Risk/scope: Medium. ASCII top-down rendering is novel (most launch carts are document-shaped) — needs a small art-side investment. Could be the cart that justifies a future BITMAP-mode push.

7. CARRION — Bounty Hunter / Skip Tracer

Pitch: Hunt fugitives. Investigation tree of leads; each lead has a confidence score; some are red herrings.
Mechanic hook: investigation tree where the gameplay is pruning (which leads to follow, which to ignore, when to commit to a lead). Resembles BLACK LEDGER’s evidence chain but with human targets and street-level legwork instead of financial trails.
KN-86 fit: Strong, but the closest existing-launch overlap of any keeper — sits adjacent to BLACK LEDGER’s investigation loop. Differentiation: CARRION’s targets are people with locations and movement patterns, not transactions.
Phase/economy fit: Bounty-payout reputation (new). Credits = bounty minus expenses. Phase: brief → investigate → apprehend / negotiate / lose.
Risk/scope: Medium. Differentiate clearly from BLACK LEDGER in the spec or it’ll feel redundant.

8. SYNAPTIC — Black-Market Chemist

Pitch: Build psychoactives molecule-by-molecule. Rings, side-chains, bonds map to s-expressions.
Mechanic hook: structural-editing on a molecular graph. Effect (the drug’s behavior) is a function of structure (the molecule). Direct mapping from “mess with the tree” to “outcome changes.”
KN-86 fit: Strong (structural editor on a tree), with a caveat — the player’s mental model has to be “molecules = expressions.” That’s a teach moment the cart needs to handle gracefully via tutorial.
Phase/economy fit: Reputation in client-base affinity (recreational vs medicinal vs exotic). Credits per batch. Phase: order → synthesize → distribute.
Risk/scope: Medium. Domain content (real or pseudo chemistry) needs a careful pass — too real reads as a how-to manual; too cartoonish breaks the device’s authenticity.

9. TACHIKOMA — Semi-autonomous AI agents (GitS)

Pitch: Command 3–5 AI agents who develop personality across sessions; each agent is a lambda you can edit.
Mechanic hook: agents-as-lambdas that the player co-authors over time. Sessions = field deployments; between-session = personality drift the player can correct, encourage, or leave alone. Conversations between agents emerge as procedural dialogue.
KN-86 fit: Strongest pure-paradigm fit in the set. The CIPHER voice is already a “competent colleague”; TACHIKOMA gives you several of them to manage. Risk: too close in vibe to CIPHER itself unless the framing distances them.
Phase/economy fit: Reputation per agent (each has its own); credits via field deployment outcomes. Phase: brief team → deploy → debrief / tune. Multi-cart: agents can run sub-missions on other carts.
Risk/scope: High writing investment per agent — each needs distinct voice + personality grammar. Sleeper pick — quietly the most ambitious paradigm play in the set.

Conditional keepers (3) — need a second pass before sprint-ready

A. HOSTILE TAKEOVER — Corporate Climber

What’s good: Org chart as a Lisp tree is on-paradigm. Verbs (blackmail, leak, golden parachute, hostile transfer) are juicy.
What’s missing: what does the player do each minute? The pitch describes the strategic level (climb the chart) but not the mechanical loop. Black Ledger has fraud trees; Ice Breaker has lattices; HOSTILE TAKEOVER’s per-minute loop isn’t yet visible. Rewrite the pitch around a single concrete operation (e.g., “build a leak: choose channels, time the release, watch the chart restructure”) before sprinting.
Recommendation: keep, schedule a 30-min Gameplay Design rev to nail the moment-to-moment loop, then re-evaluate.

B. RECITATION — Memory Smuggler

What’s good: Strange Days / Inception loop is potent; pulling specific memories out of a neural pattern is canonically cyberpunk.
What’s missing: very close to RESURRECTION’s design space (both deal in extracted/reconstructed mental content). If we ship RESURRECTION, RECITATION risks redundancy unless its mechanic is clearly distinct (e.g., RESURRECTION = assemble lambda; RECITATION = isolate fragment + transmit). Also: the buyer-driven “extract specific memory” framing has ethical heavy-lift the cart needs to handle.
Recommendation: keep but pair-design with RESURRECTION; if both ship, they must be on opposite axes (RESURRECTION = creation; RECITATION = extraction). If only one ships, RESURRECTION wins.

C. CLONESHOP — Identity Forger

What’s good: Building a fake persona piece-by-piece (birth cert, employment trail, biometric anchors) is structural-editor work and slot-fits any cyberpunk fiction.
What’s missing: the loop is assembly, which is what most other launch carts also do. CLONESHOP needs a hook that distinguishes it from CIPHER GARDEN’s authoring loop or NULL’s identity layer. Possibility: CLONESHOP outputs consume across other carts (a CLONESHOP forgery is a required input to a CARRION evasion or a HOSTILE TAKEOVER infiltration). If the cart is infrastructure for other carts, that’s its identity.
Recommendation: keep with reframe — pitch CLONESHOP as the forgery infrastructure that other carts depend on, not as a standalone game. Becomes a multi-cart-essential.

Cuts and why (3 explicit drops)

× GARGOYLE — Always-on lifelogging cart

Why cut: the device is terminal-shaped, indoor, focused-attention. GARGOYLE is wearable-shaped, outdoor, ambient-attention. The cart fights the device’s premise. The “sell timestamped footage” core loop also doesn’t translate to text-grid gameplay — what does the player do with the deck while wearing it? If the answer is “nothing,” the deck is the wrong vehicle.
Better home: a different KN-line model entirely (a future hypothetical wearable). On the KN-86 it’s wrong-device.

Why cut: “first multiplayer-essential cart” violates the device’s solo, hardware-bound, network-isolated premise. The KN-86 has Wi-Fi for system-image updates only; there’s no live network gameplay surface and no roadmap to add one without a complete arch shift. Even if we had a network surface, a social-club cart is the wrong device fit (text-grid avatars in a club don’t sell).
Better home: different product entirely (a Toneline-line social piece, maybe). On the KN-86 it’s an architectural mismatch.

× NIGHTOWL? — no, kept (mentioned only to note the close call)

NIGHTOWL nearly cut for “not on-paradigm enough” but kept because the BITMAP-mode investment it implies is good for the platform, and the location-tree shape does fit. Flagging it as the borderline keeper.

Suggested batch ordering (the first expansion wave, 5 carts)

If/when cart development goes beyond the 17 launch modules, ship these five as a coherent first wave (~6–9 month sprint, post-Q4 2027 device launch):

#	Cart	Why this slot
1	WETWARE	Clearest “you couldn’t make this on Switch” identity. Anchor cart for the wave.
2	SIGNALWAKE	Marketing surface area; instant ‘88 recognition; lowest narrative-writing cost in the set.
3	RESURRECTION	Prestige tier; the wave needs one cart that’s reviewed-as-art, not just played.
4	GHOSTLINE	Lisp-native ad demo; cross-cart compatibility with WETWARE and BLACK LEDGER opens multi-phase missions immediately.
5	DAEMON	Opens new design space (between-session play). Engineering investment now pays for many future carts.

The 4 keepers held back to wave 2 (NIGHTOWL, CARRION, SYNAPTIC, TACHIKOMA) and the 3 conditional keepers (HOSTILE TAKEOVER, RECITATION, CLONESHOP) all benefit from playtester feedback on wave 1 before specs lock — particularly on whether DAEMON’s between-session loop is teachable.

Open Questions for Josh

Are we even sprinting expansion carts post-launch? This pack assumes “yes, eventually” — if the answer is “ship 17 launch carts and stop,” GWP-283 is a research-archive task and needs no further action. Confirm intent.
GARGOYLE and BLACK SUN cuts: confirm or override. Both are good fiction; my read is they’re wrong-device. If you want either, we’d need a separate brainstorming pass on what they look like redesigned for the deck (which is real work).
CLONESHOP-as-infrastructure reframe: willing to commit CLONESHOP as a multi-cart-essential (i.e., other carts assume CLONESHOP exists)? If yes, we should pencil that into the launch-titles design bibles too — affects CARRION, HOSTILE TAKEOVER, and arguably ICE BREAKER.
TACHIKOMA scope: the agents-as-lambdas-you-edit pitch is the most paradigm-faithful in the bunch and also the most expensive to author. Worth a focused exploration session before its slot in any expansion wave is fixed.