The Execution Layer

The lane in one page — what the ten instruments read, the order they are read in, and where to start. Read this before buying any single instrument.

A deterministic conformance standard for the output of an AI decision gate. Grades each decision point ENFORCEABLE, ADVISORY ONLY or NOT A VERDICT across six elements in two groups, and ships a runnable engine, a reproducing workbook, and a working MCP reference server.

A deterministic instrument that grades whether an AI gate can actually stop an agent. Reads custody of the refusal across six observations and returns FAIL-CLOSED, FAIL-OPEN or DECORATIVE per gate, then rolls each governed tool up to its weakest gate.

A deterministic instrument for the conditions attached to conditional approvals. Grades each condition BOUND, NOTIFIED ONLY or LAPSED by testing its claimed consequence against the machinery that would deliver it, then reads each approval by its strongest condition.

A deterministic instrument that grades what each of your stop controls has been shown to do rather than what it claims. Returns STOPS, SLOWS or NAMED ONLY per control, then reads each agent by its strongest control.

A deterministic instrument for the gap between an automated action and the record of it. Grades each write-back path IN THE RECORD, PARTLY IN or UNRECONCILED by asking whether the action and its record can come apart, and whether the record is usable once it is there.

A deterministic instrument for whether an automated decision can be reconstructed by somebody who was not there. Grades the joins between the decision record and the context, policy and authorisation behind it, returning ANSWERABLE, NEEDS A WITNESS or NO RECORD.

A deterministic instrument for whether a decision already made could be produced again. Grades six sources of variance as a conjunction with no partial credit, then a three-part record as a fallback rung, returning REPLAYABLE, RECONSTRUCTABLE or UNREPLAYABLE.

A deterministic Quick Kit that grades whether an escalation arrives at somebody who can act on it, and whether their answer can change what happens. Returns REACHES SOMEONE, REACHES A QUEUE or REACHES NOBODY per path.

A deterministic Quick Kit that grades whether the verdicts you are still relying on are still good, by measuring each one's age against how fast the inputs it rests on turn over. Returns IN FORCE, AGED or VOID per verdict.

A deterministic Quick Kit that grades whether each break-glass path in your systems is accountable to a named person at the moment of use. Two dispositive answers and a five-part conjunction return ACCOUNTABLE, LOGGED ONLY or UNATTRIBUTABLE per control.

All ten instruments of the execution-layer lane, plus a deterministic roll-up engine that reads their findings in order across three stages. A later stage can never read higher than an earlier one, so the roll-up names which of your readings are moot. Returns HELD END TO END, HELD IN PARTS or NOT HELD.