• Blue line — the main health metric. For peg / aggregate runs it is the peg ratio (1.0 = healthy). For resource cascade it is minimum layer headroom. For inventory buffer it is stock level (higher = more stock). Other domains use their primary capacity metric on this axis.
• Orange line — instability (how much stress the system is under right now).
• Magenta marker — the step where the run is marked as collapsed.
• Cyan vertical line — your scrub playhead (where the mouse is on the timeline).
• Bottom strip — the shock lane: when and where the attacker injected shocks.
• Dashed lines (network replays only) — extra contagion traces (e.g. max panic, panic spread).

• Load a replay — on this public demo, use Presets or open a sample from the workbench. Local file pickers are disabled here.
• Scrub the timeline — move the mouse over the chart, or click and drag. The text panel under the chart updates with values at that step.
• Keyboard — click the chart once, then ← → to step, Home / End to jump to the start or the end.
• Compare two replays — local installs only: load a second JSON file to see deltas at the playhead.
• Choose folder (Chrome / Edge only) — local installs only.

• sample_replay.json — aggregate peg run.
• sample_network_replay.json — graph contagion (adds the dashed panic traces).
• sample_resource_cascade_replay.json — capacity-overload cascade (blue line is headroom, not a peg).
• sample_service_backlog_replay.json — operations backlog run.
• sample_liquidity_ladder_replay.json — funding ladder stress.
• sample_inventory_buffer_replay.json — stock level under demand surges and fulfillment problems.
• sample_coupled_institution_replay.json — research fork: blue = peg price, orange = instability, teal = overload (coupled in one step()).

All samples are reproducible from frozen seeds; the panel under the chart shows the exact CLI that produced the file.