Context

We've decided on the "Context Continuity as shared coordination state" framing (@Agent prompt reframing: Context Continuity as shared coordination state) and the strategic direction of Native as a coordination substrate (@Strategic direction: Native as coordination substrate (land-and-expand)). Now we need to figure out, from first principles, what this framing means for every aspect of how we guide agents.

What This Task Covers

Start from the mental model ("the workspace is shared state that multiple agents and humans read from and write to") and work out:

1. What are all the things agents do in the workspace? Don't take the current set of guides/behaviors as given. Think from scratch about what matters for coordination.

2. For each behavior, what principle flows from the coordination framing? e.g. Why should agents link records? Why place things in hierarchy? Should they? What about archiving, titling, status, decisions, tags, mentions?

3. What's the minimal set of principles that covers everything? Avoid a long laundry list — find the small number of principles that generate the right behavior across all cases.

4. Where does each principle live? Bootstrap preamble vs. bootstrap principles vs. guides vs. prompt workflows. What's the right layering?

Important

• Don't take current guides (placement, hierarchy, context-first) as granted. They were written with a different framing. Some may survive, some may merge, some may not be needed, new ones may emerge.
• Think about what an agent actually needs to know to coordinate well, not what we've historically told them.
• Consider edge cases: what happens when agents over-structure? Under-structure? What's the failure mode of each principle?

Output

A documented set of principles with rationale, organized by where they should live (bootstrap vs. guides), ready to inform the implementation tasks for the 5 file rewrites.