Skip to content

fix(mobile): preserve grouped project workspaces - #4642

Merged
shivamhwp merged 30 commits into
pingdotgg:mainfrom
shivamhwp:agent/fix-mobile-project-grouping
Aug 5, 2026
Merged

shivamhwp merged 30 commits into
pingdotgg:mainfrom
shivamhwp:agent/fix-mobile-project-grouping

Merge branch 'main' into agent/fix-mobile-project-grouping

db0db83
Select commit
Loading
Failed to load commit list.
MacroscopeApp / Macroscope - Effect Service Conventions succeeded Aug 5, 2026 in 5s

All clear

All clear

Details

Note

Your check run agent prompt is: .macroscope/check-run-agents/effect-service-conventions.md
More information about how Check Run Agents work can be found in our Help Center.


🔍 Agent activity — reasoning, no tool calls

💭 Looking at this PR, I need to check whether it's creating or consuming an Effect service. The changes span project grouping logic across the client-runtime and mobile app—mostly pure functions in the shared package, with some hook usage in the mobile layer. Let me examine the specific files to see if there are any Effect service conventions being violated. Looking at the diff for Effect service convention violations, I'm checking whether the re-export pattern in the project-grouping module violates the rule about removing compatibility shims. The convention states not to leave re-export shims when refactoring, but this appears to be a hook module rather than a service layer, so it might not apply here. I'm also verifying the imports—there's an AsyncResult import from the unstable reactivity module which follows the repo's pattern, and some imports like pipe and Arr were removed from new-task-flow-provider, though I'd need to confirm whether they're actually unused in the file or if that's a separate linting concern outside the scope of this check. Looking at the projectGrouping module, it's just pure functions with no services involved. The scopedProjectKey calls use different signatures across modules but they're from different imports, so that's expected. I don't see any Layer.succeed or service injection patterns in the diff, and the hook consumption in NewTaskRouteScreen doesn't require Effect provider setup. I'm fairly confident there are no Effect service convention violations here.