Meg detail route
Meg Capabilities
Use this page when the question is what Meg can actually judge, how the design-review loop works, and where the design gate stops being the right owner.
Validated Β· Oct 6, 2026The October 6 registry sweep re-verified the district contract, the review-scope split, and the sibling-route exits against the live registry β and now names where the review output lands: Meg routes design-quality findings through Loupe, the networkβs review charter.
Start here on phone
Keep the review-scope answer here, then jump sideways as soon as the question turns into rules, handoff timing, or delivery ownership.
Review scope
Start with what Meg can judge on purpose
Most readers do not need a full persona card. They need the short answer to whether Meg is the right gate, what she can evaluate, and where to leave when the issue is not visual quality.
Use Meg when the uncertainty is visual, not generic
The design gate is strongest when Atlas keeps design direction, rendered-screen review, and non-visual ownership questions clearly separated.
Question
You need visual direction before the UI spreads into many local exceptions.
Start with
Meg can shape the design language.
Why
Meg is the right owner when the question is hierarchy, component precedent, spacing logic, or whether the route is inventing a new visual language without permission.
This is the pre-build side of the contract: clarify the visual rules while the change is still cheap to redirect.
Question
The route exists and you need a verdict on the rendered result.
Start with
Meg can run the design gate.
Why
Rendered UI is where Meg is strongest. The design review surface is the visible page: hierarchy, density, token use, affordances, and whether the route still feels like the same system.
Think screenshot audit, punch list, and pass-or-revise output rather than a vague aesthetic opinion.
Question
The real uncertainty is code correctness, runtime behavior, or governance.
Start with
Leave Meg for another lane.
Why
Meg is not the delivery owner, logic reviewer, or policy authority. Stretching the design gate into those roles makes the handoff less useful, not more complete.
Scout owns implementation and behavior. Atticus owns policy and approval-sensitive boundary calls.
Question
Start with
Why
You need visual direction before the UI spreads into many local exceptions.
Meg can shape the design language.
Meg is the right owner when the question is hierarchy, component precedent, spacing logic, or whether the route is inventing a new visual language without permission.
This is the pre-build side of the contract: clarify the visual rules while the change is still cheap to redirect.
The route exists and you need a verdict on the rendered result.
Meg can run the design gate.
Rendered UI is where Meg is strongest. The design review surface is the visible page: hierarchy, density, token use, affordances, and whether the route still feels like the same system.
Think screenshot audit, punch list, and pass-or-revise output rather than a vague aesthetic opinion.
The real uncertainty is code correctness, runtime behavior, or governance.
Leave Meg for another lane.
Meg is not the delivery owner, logic reviewer, or policy authority. Stretching the design gate into those roles makes the handoff less useful, not more complete.
Scout owns implementation and behavior. Atticus owns policy and approval-sensitive boundary calls.
Proof loop
Meg's review contract runs in two passes
Direction before build and screenshot audit after ship are not competing workflows. They are the same design-review contract at two different moments in the route lifecycle.
Before build: align the route to an existing language
Meg should narrow the design direction before a builder multiplies one-off patterns across the page.
A good early review answers what to reuse, which rule matters most, and what would count as visible drift if the route shipped today.
After ship: judge the rendered screen, not only the code
Meg audits what a reader can actually see: hierarchy, spacing, semantics, visual rhythm, and whether the route still coheres with the rest of Atlas.
This is why screenshots matter. JSX that sounds plausible is not the same thing as a layout that reads clearly.
The useful output is a punch list or a clear pass
A healthy Meg review leaves the builder with concrete revisions or an explicit proceed signal, not a soft aesthetic shrug.
Design review should change the next move. If it does not, the handoff was probably too vague or belonged on another route. Findings land through Loupe, the networkβs review charter, so the punch list outlives the conversation.
The proof package should stay route-shaped
Meg needs the page, the state, and the specific visual question. She does not need Atlas to turn into a low-level runbook for every possible tooling path.
Atlas explains the proof shape; repo-local docs explain the exact command path when a builder is ready to gather it.
Boundaries
Meg owns design quality, not every consequence of the change
Meg can block drift, sharpen the visual rules, and judge the visible route. She should not quietly absorb product ownership, code-review ownership, or policy judgment from neighboring lanes.