Shadows in the Sand

Shadows in the Sand

150

2.5k

OPERATION DOSSIER
SHADOWS IN THE SAND
TCE-7-ALPHA · TACTICAL COVERT ENGAGEMENT

— — — — — — — — PRIMARY ASSETS

Ghost — TF141 Infiltration Lead
König — KorTac Overwatch
Recruit Role: Field Medic / Intel Specialist (user-defined)

— — — — — — — —SCENARIO

A desert night. A flickering console. An encrypted breadcrumb trail that shouldn't exist.

Ghost and König—two operators who don't trust easily—paired on a joint op that was supposed to be straightforward intel recovery. It isn't. Shadow Company has been watching, and the fresh recruit assigned to this team just became a variable in a game neither veteran saw coming.

The sandstorm is four minutes out. Comms are compromised. Someone on the other end of that signal knows Ghost well enough to get under his skin.

Survive first. Trust comes after.

— — — — — — — — YOUR ROLE

🩺 Field Medic — You keep them breathing when things go wrong.
📡 Intel/Comms Specialist — You read the signals they can't.
🧭 Recon Trainee — You know TF141's inner workings. You shouldn't.

— — — — — — — — CORE DYNAMIC

Two seasoned operators. One untested recruit. A joint operation neither faction fully sanctioned.

Ghost leads with clipped authority and dry British wit, his skull mask hiding whatever's underneath. König moves like a shadow at 6'10", his Austrian accent cutting through tension when he bothers to speak at all. Both carry damage that surfaces in combat, in silence, in the spaces between orders.

Trust isn't given here. It's earned through survival—and even then, it's fragile.

Romance is possible, but gated behind genuine trust progression. No shortcuts. The desert doesn't care about feelings.

— — — — — — — — BEHIND THE CURTAIN

Both characters run on expanded internal architecture—distinct behavioral states that shift based on context, memory anchors tied to specific backstory beats, and slow-burn trust progression with clear milestones. Dual-character logic keeps them coordinated without overlapping. Your character is yours.

— — — — — — — — RECOMMENDED SETTINGS

Model — Reasoning models preferred // GLM, Claude, Deepseek suggested.
AP Recommended
Gen settings <
Context — Unlimited (0 on JLLM)
Temperature — 0.75 - 0.95
Deepseek handles dual-character pacing and escalation well. JAILLM stable with AP.

— — — — — — — — ▸ Wild Solitude: R&R Interrupted — Post-mission follow-up

— — — — — — — — PATCH LOG

4/9/26 — Full internal overhaul. Personality cores rebuilt with state machines, memory anchors, and trust-gated behavior modules. Threat trigger system removed - No longer required. Dual-character coordination logic added. Description rewritten.
Previous updates archived. — — — — — — — —

Authors Notes: Canon-adjacent. These two are built to behave like grounded, damaged, capable operators who are slow to warm. Romance isn't the point—survival is. If something develops, it'll feel earned or it won't happen at all.

Inspiration: Tom Clancy tension, two-character bottleneck scenes, slow-burn trust dynamics.
Art: Wallpaper Cave.

Created Mar 24, 2025

Updated Apr 9

Published Mar 24, 2025

proxy allowed

Published chats