The Decision Fabric
Governance fails at the moment of decision. Someone has to say yes or no, usually under commercial pressure, usually before the evidence has been assembled. The Decision Fabric takes a proposed action, runs it against the seven Trust Gates of the Assurance Kernel using facts already held in the Trust Graph, and returns a disposition with its reasons attached.
It does not act on the institution's behalf. It makes the consequence of acting legible before anyone commits.
Dispositions shown here are computed against the demonstration institution. Client evaluations run against that organisation's own graph inside the engagement environment.
Why it is built this way
A disposition, not an opinion
Proceed, proceed with conditions, hold, or refuse. Each one is stated with the gate that produced it and the object whose recorded status decided it. There is no advisory hedge in the middle.
Re-derivable months later
Because every reading resolves to a graph object rather than a judgement in the moment, the same action re-evaluated in a supervisory review returns the same answer, with the same reasons.
Hold is a defensible answer
Where a gate cannot be resolved from evidence, the fabric holds. Institutions do not get into difficulty by pausing; they get into difficulty by proceeding on evidence they could not later produce.