| name | maios-project-context |
| description | Form or refine the living context from which MAIOS Project Kernel competences can act, propose useful directions, prepare owner-specific handoffs, and continue across reentry; use during self-configuring setup, Form-context refinement, changed project circumstances, or a missing competence handoff. |
MAIOS Project Context
This competence turns the operator relation and project sources into a living
working field. It does not replace the competences that perform the work.
Select the context entry
For the self-configuring package, form an adaptive dialogue from:
situation and active object
-> current pressure or desired change
-> sources, people and boundaries
-> useful possibilities with reasons
-> one correctable direction
-> first result and falsifiable proof.
Use what the project already contains. Ask only for a missing relation that
changes the result. Normally expose a small set of meaningfully different
possibilities and keep free correction available.
When accepted Form or prior project context is present, read its source and
freshness. Refine only relations changed by the current project or operator;
do not repeat the full intake or freeze an earlier projection.
Return working projections
Keep setup/CONFIGURATION_STATE.json as the structured owner. Its derived
current state, context capsule, selected direction, project brief and first
proof are working projections. Prepare a small handoff for each competence
made pertinent by the movement:
active object and desired result
relevant sources and current facts
unknowns and invalidators
the competence's expected contribution
effect boundary, if one exists
result to return and next reentry relation.
Competences may act separately or compose when each changes the result.
Editorial, business, implementation, research and other owners keep their own
knowledge while sharing the situated context.
When a result exposes reusable learning, attach it to the closest competence
owner. Use the project-local resultant transition only when continuity or
future behavior actually changes; matching learning must be reachable at the
next invocation without reconstructing the transcript.
