APEX-BG

APEX-BG was a behavioural scoring engine that qualified an agent before it was allowed to run. In the owner's words, it was one of the "neat tricks developed before the industry really got there."

That exact.works scored agents on three gates — behavioural fitness, consistency, reliability — kept a permanent behavioural record, and blocked agents that failed to qualify. Scoring an agent means holding a record about it and gating on that record. That is operating a platform and enforcing against its participants, and exact.works does neither.

Nothing here. NeMo Guardrails, Azure AI Content Safety and LangSmith evals own this ground now. That is not a gap worth re-entering.

exact.works helps two parties write down what an agent may and may not do, and compiles those agreed terms into machine-readable configuration for whichever stack they already run. It specifies; it does not score, gate or supervise. The scoring methodology page is withdrawn on the same terms.

Nothing. It was withdrawn rather than replaced.