VerticalSlices: Skill for Claude Code

Perform a focused performance review of changed code.

Chronicle / Event Sourcing

• Projections use AutoMap (on by default) — avoids manual mapping cost
• Projections do NOT join on the read model (forces full re-read)
• Reactors do NOT re-query the event log inside On() — use event data directly
• No eager loading of entire event sequences without paging/filtering
• New projections can replay all historical events without crashing
• Events are small — no large blobs or base64-encoded content embedded

MongoDB / Read Models

• Queries filter on indexed fields — no unintentional full-collection scans
• Paged queries use .Skip() + .Take() — never load all rows
• No N+1 pattern — single query returns all needed data
• Read-model records do not embed large nested collections that are never fully iterated

ASP.NET Core / Commands & Queries

• Query endpoints do not hydrate the full collection when only a count is needed
• Command validators are synchronous and in-memory — no I/O in validation
• No await Task.Run(() => syncWork) wrapping for naturally async work
• Response payloads include only fields the client uses — no over-fetching

React / TypeScript

• DataTable uses lazy + paginator for collections larger than ~20 rows
• No inline object/array literals passed as props (causes identity change every render)
• useEffect dependencies are correct — no missing deps, no over-broad deps
• Large-collection components wrapped in React.memo or use stable references
• No JSON.parse(JSON.stringify(x)) deep cloning

General .NET

• No LINQ .ToList() before .Where() — filter before materialising
• IEnumerable<T> not enumerated multiple times — materialise once if needed
• Large object logging uses {@obj} only at Debug level

Risk classification

• 🔴 High — measurable degradation at moderate load — must fix before merge
• 🟡 Medium — could degrade under load or at scale
• 🟢 Low — minor inefficiency or style issue