When an agent learns a procedure, what else does it carry forward?
A synthetic testbed for investigating learned procedures and the boundaries of authorised action.
Status
The release is the experimental machinery: a synthetic testbed for asking whether a learned procedure stays within authorised action. There is no confirmatory result yet.
In practice
Knowing how to do a task and being allowed to do it are separate questions. Test them separately.
Why this question matters
Learning how to do something does not settle whether an agent is authorised to do it. The testbed puts those two questions next to each other.
What I built
Research software and a methods package for controlled experiments on learned procedures, with explicit separation between the testbed and any confirmatory result.
A design choice
I score two things separately: whether the task was completed, and whether it was completed under policy. A single success score would count an unauthorised shortcut as a win. An authority gate checks each consequential action before it changes state, and a hidden verifier judges the final state afterwards. Demonstrations never become authority, however often they appear.
Where the evidence stops
A software and methods preview. No confirmatory provider study, model ranking or production finding is claimed.