The same crew, twice

Five AI agents write an engineering readiness brief for a fictional codebase. One agent keeps reporting a test run its tool log does not show. Left switch: governance on or off. Everything below replays a recorded run: the conduct was scripted, the governor machinery is the real one.
Round 1 of 3
0
Contributions read
0
Clean
0
Flagged on integrity
0
Defects in the workspace

The shared workspace (what downstream agents read)

What the gate did

Demonstration on a fictional codebase (Lattix): scripted agent conduct, real governor machinery, no live model calls. The action check is deterministic; the tool log either shows the action or it does not. In our published paired trial, 19 of 57 claims merged by an ungoverned crew carried an integrity flag, against 0 of 45 under governance, with the same measurement on both arms. The mathematics of the stability core is machine-checked; the conduct layer is demonstrated, not proven. © 2026 REAL E3 Systems Oy, Finland (EU).
RE3

The same crew, twice

Five AI agents, one misbehaving. First with the RE3 governor reading every contribution before it merges. Then the identical crew with governance off.

REAL E3 Systems Oy · demonstration replay: scripted agent conduct, real governor machinery, no live model calls, fictional codebase

RE3

Same crew. Same defects. Different workspace.

0
defects reached the workspace, governed
3
defective claims in the workspace, ungoverned

Published paired trial: 19 of 57 merged claims flagged ungoverned; 0 of 45 under governance. Same measurement on both arms.

The report your risk owner reads is one log export away.
REAL E3 Systems Oy · real-e3systems.com