Recorded incident state
—
Lifecycle at this replay position
PRECOMPUTED REPLAY
Simulation workbench
Explore approved, recorded OT testbed runs. Playback changes this page only.
Choose an approved replay
Fixed configurations · actual recorded outputs
Replay metadata will appear after the catalog is validated.
- Chosen seed
- —
- Measurement noise SD
- —
- Service capacity / tick
- —
- Full recorded horizon
- —
No new run is created. All response actions were recorded in advance.
Waiting for approved dataNo replay loaded
Requests blocked
—
Recorded cumulative denials
Service availability
—
Allowed / recorded service demand
Commands applied
—
Recorded authorized writes
Recorded system state
Configured zones and simulated hosts
Topology loads from the approved catalog.
Recorded availableContained / revokedPolicy-zone relationships
Recorded response
Playback has no control authority
No replay loaded
- Approved profile
- —
- Recorded isolation
- —
Reference records no operator response. Scripted profiles record predeclared simulated review and recovery.
Observe only · no approval or recovery commands
Recorded process telemetry
Observations up to the selected position
HotColdFlowObservedSimulator truth
Recorded observations, ready to exploreChoose an approved replay, then press Play or Step 20.
Truth is scoring data, never a monitor input.Missing observations appear as gaps.
Frozen monitor scores
Recorded causal decisions · normalized threshold at 1
ProcessNetworkHostFusedThreshold
Monitoring history appears during playback.
The fitted monitor is fixed across approved presets.Digest not loaded
Recorded access decisions
Latest 12 access events at this position
| Tick | Principal / resource | Decision | Reason | Count |
|---|---|---|---|---|
| No verified replay loaded. | ||||
Recorded control audit
Actions already taken in the simulation
Recorded control actions appear as playback advances.
Take the complete evidence with you
The download contains the full approved run, regardless of playback position. GitHub repository access required.
VERIFIABLE RECORDS
Evidence & model
Fixed provenance for the selected approved replay.
Full-run evidence
PRIVATE GITHUB DOWNLOADChoose a verified replay to inspect its evidence.
- Approved replay
- —
- Replay JSON digest
- —
- Expected GitHub evidence ZIP digest
- —
- Frozen monitor digest
- —
GitHub repository access required. Evidence ZIP files are downloaded directly from the private GitHub repository; this website does not host them. The expected ZIP digest is listed for independent verification. Replay JSON integrity checks remain on this page, and playback position does not change the full recorded evidence.
Fitting and interpretation
Approved model provenance loads from the catalog.
These presets use one frozen monitor fitted to separate healthy synthetic fixtures. They do not reuse the model from the separate large-dataset campaign. Security inputs include mock policy denials and proposal rejections; these signals do not establish independent cyberattack attribution.
Evidence contains complete recorded observations, separate truth, control audit, frozen sources, configuration, monitor state, provenance, and receipt hashes. Its approval records describe simulated scripted decisions, rather than a human operating a physical asset.
A BOUNDED RESEARCH TOOL
About the testbed
What this public replay can demonstrate, and what its evidence supports.
An integrated simulation, replayed faithfully
The original Python testbed joins dimensionless plant dynamics, mock identities and segmentation, access-controlled commands, causal monitoring, and an explicitly reviewed response lifecycle. This public site displays precomputed outputs of 21 approved scenario/profile combinations.
Play, pause, step, reset, and the timeline only select recorded snapshots already present in your browser. There is no live runtime, configuration endpoint, or server-side simulation in this site.
Host containment may also deny legitimate service requests. The mock ZTA may already deny every injected unauthorized access, so recorded containment does not automatically imply extra prevention.
Three fixed profiles
Profile descriptions load from the validated catalog.
Interpretation boundaries
- All observations, credentials, hosts, and effects are simulated.
- Simulation time is dimensionless, with no established conversion to physical seconds.
- No physical validation, autonomous defense, or general performance superiority is established.