DPL — Decision Plane Laboratory
An independent, deterministic simulation of a decision plane. It shows
how a policy takes a short path, evaluates further, asks for
evidence, waits for approval, abstains or blocks an action. No
scenario action is actually executed.
- Six versioned scenarios, from clear tasks to untrusted content.
- Three decision policies: Fast-first, Deep-first and Adaptive.
- An ordered gate contract with a step budget, abstention and blocking.
- Evidence requests, human approval and a readable run trace.
- Side-by-side policy comparison and JSON run export.
This interactive laboratory needs JavaScript. Its scenarios, policies
and results are educational design choices, not validated measurements.