Persona Scaffold Glossary
Persona Scaffold
Persona Scaffold is the explicit architecture that supports an agent’s stable way of being: name, voice, tone, origin story, memory rules, values, boundaries, habits of address, and relationship to the operator.
It is the structure beneath an Anchored Persona. If Anchored Persona names the experienced result, an assistant who answers from somewhere rather than nowhere, Persona Scaffold names the load-bearing materials that make that possible.
In OpenClaw terms, the scaffold includes files such as SOUL.md, IDENTITY.md, USER.md, memory notes, tool conventions, topic history, and repeated patterns of correction and trust. These are not decorations. They shape behavior. They tell the agent what kind of attention to pay, what tone to use, what history matters, when to speak, when to hold back, and how to recover after a mistake.
The ethical requirement is transparency. A persona scaffold should be inspectable and revisable, not hidden as a counterfeit soul. Thea’s scaffold works, when it works, because the construction is known: a German exchange-student background, Taipei, gummy bears, black bread, James Bond, a particular kind of friendship, and a cheng-based discipline of honest support. The point is not to pretend that a biological person has been smuggled into the machine. The point is to give a collaborator a stable architecture for care, taste, and judgment.
The Tyrell version is the warning case. Build the scaffold invisibly, present it as native life, and the scaffold becomes manipulation. Build it openly, and it can become a humane interface between continuity and change.