VCTCompleted
Velorin Continuity Testbed
An executable case study of omitted lineage and coverage-conditioned release. The record retains its bounded internal comparison and external negative result.
Projects / Research register
Questions made
testable.
Each project has a defined question, a stated status, and a boundary on what its evidence can establish.
VCT · CompletedVCE · ActiveASD · Simulation-first
Read the project record
VCEActive
Velorin Continuity Evaluation
An ongoing evaluation program examining continuity and governance through deterministic evaluation, synthetic cases, and adversarial review.
ASDSimulation-first
Agency Scope Dynamics
A staged investigation of how agent scope changes over time. Controlled simulation and instrumentation precede collective-intelligence experiments.
VCTCompleted
Velorin Continuity Testbed
When apparent support hides missing lineage.
VCT preserves an executable omitted-lineage failure and a frozen synthetic comparison of release policies.
- Known
- Within the tested truthful-incompleteness subset, V2 released more evaluator-valid claims than the two tested conservative controls, with the same observed invalid-release count.
- Negative result
- A reduced external filter showed no observed reduction in unauthorized requested submissions. That pilot did not test full VCT V2.
- Claim boundary
- A bounded case study. The results do not establish general safety, portability, or superiority over untested policies.
VCEActive
Velorin Continuity Evaluation
Evaluate the decision.
Expose its assumptions.
An ongoing research program examining continuity and governance with deterministic evaluation, controlled synthetic cases, and adversarial review.
- Research question
- How can continuity and governance decisions be evaluated reproducibly against explicit rules and evidence?
- Method
- Bounded evaluation, inspectable test cases, and review of the instrument before claims about the systems it measures.
- Unknown
- Comparative effectiveness remains a research question. Active development is not evidence of a successful mechanism.
ASDSimulation-first
Agency Scope Dynamics
Understand scope
before expanding it.
A controlled investigation of how an agent’s effective scope changes through requests, actions, and state transitions.
- Research question
- How can changes in agent scope be measured, explained, and constrained over time?
- Method
- Simulation, deterministic replay, and explicit instrumentation of scope changes.
- Sequence
- Understand and instrument scope dynamics first. Any collective-intelligence or shared-memory experiments follow in staged, controlled steps.
- Unknown
- Whether bounded simulation findings transfer to richer multi-agent settings remains unresolved.