Scenario & Conditions
Pick a scenario and toggle failure modes. The map will highlight how the agent moves through its state machine.
Agent triages incidents from alerts, gathers context, and escalates to the on-call engineer with a summary.
Failure & Drift Toggles
Turn on different failure modes to see how the agent moves into error, timeout, or fallback states.
State Map & Trace
No trace yet. Choose a scenario, toggle conditions, then click Run State Trace.
State Transition Skeleton
States are fixed; the highlighted path shows how this specific scenario flows.
Initial State
Task received, context assembled, rules loaded.
Next / Processing State
Agent is executing the task under current conditions and rules.
Timeout State
Dependencies are too slow or unresponsive; work is blocked.
Error State
Agent detects a hard failure: missing inputs, conflicts, or invalid path.
Fallback Handler
Recovery logic: retries, degraded modes, escalation, or human handoff.
Recovery / Retry
Agent attempts to continue with adjusted inputs, reduced scope, or new strategy.
End State — Normal Completion
Task finished as intended; no major drift or failure.
End State — Degraded Success
Task finished with caveats (fallback, partial output, or manual intervention).
End State — Failed / Stalled
Agent could not safely complete the task; requires redesign or debugging.
Bright highlight = final state for this scenario.
Soft outline = states visited along the way.
State Trace
This log translates the path into plain language — useful for debugging and interviews.
- No trace yet. When you run a scenario, each state transition will appear here in order.
Interpretation Notes
Once you run a scenario, this section will summarize how the selected conditions shaped the agent’s path — e.g., “Timeouts forced a fallback branch, but the agent still reached a degraded end state.”