perf-review
Skill214 starsChanged 30 days agoMITView on GitHub
Performance impact assessment for items with the needs-perf-review trait. Evaluates hot paths, query patterns, and measurement plans. Invoked via skillPointer when filling performance-baseline notes.
4 other projects use a version of this.
Adopted by
Most recently taken up by
- vhmarquez/argotheir own version, first seen Oct 15, 2026
- octanejs/octanetheir own version, first seen Oct 9, 2026
- DataDog/dd-trace-javatheir own version, first seen Oct 8, 2026
- Seriora-Research/seri-agenttheir own version, first seen Sep 5, 2026
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- octanejs/octane★ 1.5ksame name
- DataDog/dd-trace-java★ 739same name
- vhmarquez/argo★ 1same name
- Seriora-Research/seri-agent★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository12 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- claude-plugins/CLAUDE.mdCLAUDE.md
- add-componentSkill
- api-compat-reviewSkill
- feature-implementationSkill
- implementSkill
- migration-reviewSkill
- plugin-impact-reviewSkill
- prepare-releaseSkill
- review-qualitySkill
- spec-qualitySkill
- test-authorSkill
The file
Files
Performance Review Framework
Evaluate performance impact of changes. This project is a Kotlin MCP server with SQLite via Exposed ORM, handling tool calls synchronously per request.
Step 1: Hot Path Analysis
Identify which hot paths the change touches:
- Per-request paths — MCP tool execution (every tool call hits this). New work here adds latency to every request.
- Per-item loops — operations that iterate over items (search, overview, stalled-item detection). N+1 patterns here scale poorly.
- Startup path — server initialization, Flyway migrate, data steps, the FTS integrity check, server-config import. Affects container startup time.
- Background operations — cascade detection, dependency resolution. Runs inline, not async.
Step 2: Database Query Patterns
- N+1 queries — does the change add a query inside a loop? (e.g.,
countChildrenByRoleper child in overview). Count total queries for a typical operation. - Full table scans — any
selectAll()without filters on large tables? - Missing indexes — new filter conditions that would benefit from an index?
- Transaction scope — are transactions held open longer than necessary?
- Aggregate vs fetch-all — using
SELECT COUNT(*)withGROUP BYvs fetching all rows and counting in memory?
Step 3: JSON/Serialization Cost
- Large response payloads — does the change add fields that significantly increase response size? (e.g., adding
childCountsto every child in overview) - Repeated serialization — same object serialized multiple times in one request?
- Trait hydration — item traits are rows in
item_traits, attached to items by one extra batched statement per read (WorkItemRows.withTraits). Flag a read path that loads items per row instead, or one that skips hydration.
Step 4: Complexity Analysis
- What is N? — identify the scaling variable (number of items, children, notes, dependencies)
- Current complexity — O(1), O(N), O(N*M)? Where does the change sit?
- Realistic scale — what's the expected N in practice? (Most projects: <100 items, <30 children per root)
- Worst case — what happens at 1000+ items? Does it degrade gracefully or hit a wall?
Step 5: Measurement Plan
- How to verify — what should be measured before/after? (query count, response time, payload size)
- Baseline — document current performance for the affected operation
- Acceptance threshold — what's the maximum acceptable degradation?
Output
Compose the performance-baseline note with findings from each step. This note is optional (required: false) — use it when the change touches known hot paths or adds significant new work.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

