Welcome to Open Questions — the problems the framework has not yet solved

[vdmeu]

The framework is honest about what it does not yet know. These are the open questions any contributor can engage with.

1. Live provenance data model
What does a continuous provenance signal actually look like technically? What events generate a signal? What is the minimum data model?

2. Consent and GDPR
Competence provenance data is personal data. How does Cupel handle consent, data portability, and GDPR compliance while still delivering useful signals to organisations?

3. Gaming and Goodhart's Law
Any provenance system becomes gameable once people know what signals are being measured. What makes the multi-signal model specifically resistant? What attacks remain feasible?

4. Primary stakeholder
Is the primary stakeholder the individual whose competence is being tracked, or the organisation that needs to trust them? These interests are not always aligned. What architecture serves both?

5. Below the abstraction layer — operationalisation
Cupel's differentiation depends on verifying competence below the AI abstraction layer. What does a concrete test of this look like in a specific domain?

Start a new thread on any of these. Disagreement is as useful as agreement — the goal is a defensible answer, not a comfortable one.