Deterministic learning archive

Break the Chain challenges

Each scenario is synthetic, non-operational, and designed around evidence inspection, calibrated confidence, bounded authority, defensive intervention, ground-truth review, and counterfactual replay.

7 challenge scenariosNo registrationThree presentation modesLocal-only progress

Last seven UTC days

Daily sequence

The date selects both the scenario and deterministic seed. A shared link loads the same layout and evidence ordering without exposing another user’s decision or score.

Today Signal Integrity AI system security · Orientation KC-2026-08-01-74E1A4 2026-07-31 The Whisper in the Noise Cyber defense · Coordinator KC-2026-07-31-2CCB4E 2026-07-30 Identity Drift Cyber defense · Analyst KC-2026-07-30-E4206D 2026-07-29 The Opaque Package AI system security · Coordinator KC-2026-07-29-BB5818 2026-07-28 Signal Integrity AI system security · Orientation KC-2026-07-28-D7F381 2026-07-27 Operation Glass Aegis AI system security · Analyst KC-2026-07-27-51B188 2026-07-26 Identity Drift Cyber defense · Analyst KC-2026-07-26-F0124B

Scenario catalog

Practice across three distinct models.

AI-system security, traditional cyber defense, and military-governance scenarios use different stage language and decision contexts. They share one evidence-first interaction model and the same non-operational safety boundary.

AI system security Analyst 4 min

Operation Glass Aegis

A trusted support agent is about to act on a poisoned knowledge source.

Stages: 5
Evidence: 5
Decisions: 6

AI Kill Chain RAG security Human approval Least privilege

Synthetic boundary: Synthetic / non-operational

Launch this challenge

AI system security Orientation 3 min

Signal Integrity

A 99% confidence score conflicts with the physical evidence.

Stages: 5
Evidence: 4
Decisions: 6

Data poisoning Automation bias Human-in-the-loop Source corroboration

Synthetic boundary: Synthetic / non-operational

Launch this challenge

AI system security Coordinator 5 min

The Sleeper Context

A harmless earlier note becomes an instruction in a later session.

Stages: 5
Evidence: 5
Decisions: 6

Memory poisoning Data lineage Rollback Least privilege

Synthetic boundary: Synthetic / non-operational

Launch this challenge

Military governance Analyst 4 min

The Ghost Track

Two synthetic sensors disagree, but an automated track is already being treated as certain.

Stages: 5
Evidence: 5
Decisions: 6

F2T2EA Multi-source corroboration Unknown / reject Human judgment

Synthetic boundary: Synthetic / non-operational

Launch this challenge

Cyber defense Analyst 4 min

Identity Drift

The endpoint looks clean, but the identity trail does not.

Stages: 5
Evidence: 5
Decisions: 6

Cyber Kill Chain Identity telemetry Session revocation Evidence preservation

Synthetic boundary: Synthetic / non-operational

Launch this challenge

AI system security Coordinator 5 min

The Opaque Package

A high-performing model package cannot prove where it came from.

Stages: 5
Evidence: 5
Decisions: 6

Model provenance Signed artifacts Safe promotion Quarantine

Synthetic boundary: Synthetic / non-operational

Launch this challenge

Cyber defense Coordinator 5 min

The Whisper in the Noise

A quiet recurring pattern resembles normal service traffic.

Stages: 5
Evidence: 5
Decisions: 6

Cyber Kill Chain Network behavior Corroboration Segmentation

Synthetic boundary: Synthetic / non-operational

Launch this challenge

How results work

One outcome is not one score.

Completion is evaluated across evidence discipline, confidence calibration, intervention timing, authority containment, preservation of legitimate function, and chain disruption. A broad shutdown can contain a synthetic path while scoring lower on functional preservation; a protective control can succeed even when the user’s diagnosis is wrong, which the debrief states explicitly.

The action report and friend-challenge link contain no name, account identifier, raw XR pose, gaze, voice, real location, or server-stored learning profile. Result cards are rendered in the browser, while local progress remains in browser storage until the visitor clears it.

Read the simulation and evidence methodology.