Performance work as measurement discipline — baseline first, one change at a time, beat the noise, and a keep/revert verdict where neutral is a revert. INVOKE when making anything faster or lighter, investigating slowness, or reviewing an optimization — "optimize", "too slow", "reduce latency/bundle size" — even mid-build. NOT for diagnosing functional bugs ([[debugging]]) or query-shape hygiene while writing ordinary code ([[code-standards]] §5).
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# Performance — the measurement is the work
Most "optimizations" are guesses wearing confidence: a change lands because it *should* be faster, nobody measures, and the codebase accretes complexity that never bought anything. The discipline that prevents this is not knowing the tricks — it's the loop around them: measure a baseline, change one thing, re-measure the same way, and **keep only what provably paid**. Code you keep, you maintain forever; make it pay for itself.
## How to apply this
Run the loop in order: baseline → identify the bottleneck → one change → verify against the baseline → keep or revert → record the attempt. When invoked directly as `/sonu:performance`, apply it to `$ARGUMENTS` — the text typed after the invocation; if that token appears literally or is empty, apply it to the performance concern in the current discussion.
---
## 1. Baseline before touching anything
…
The line at the top of this page pins the label git:20260826.d0dd268, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
GET https://markdownregistry.com/api/v1/artifacts/art_webetdr7h3wulczf
GET https://markdownregistry.com/api/v1/resolve?ref=prabhdeepsingh/claude-plugins/performance
GET https://markdownregistry.com/api/v1/blob/8a6469143afc0a637e7375c77ae0adfbcaf8c974f3eb2ffaff1937a690f82014
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
prabhdeepsingh/claude-plugins · .claude/skills/plugin-dev/SKILL.md · The maintainer's handbook for the sonu plugin and the claude-plugins repo. INVOKE whenever authoring, editing…
prabhdeepsingh/claude-plugins · sonu/skills/blast-radius/SKILL.md · Consumer-impact discipline for contract changes — enumerate every consumer, flag the ones that degrade silently, verify…
prabhdeepsingh/claude-plugins · sonu/skills/bug-finder/SKILL.md · Hunt for one real, previously unreported defect and file it as a well-evidenced ticket — proactive discovery, not…
prabhdeepsingh/claude-plugins · sonu/skills/classify-tickets/SKILL.md · Backlog hygiene as a sweep — exactly one type and one evidence-based priority per open ticket, and nothing else…
prabhdeepsingh/claude-plugins · sonu/skills/code-standards/SKILL.md · Prabhdeep (Sonu) Singh's personal coding standards — the house rules and quality bar for how code gets written. INVOKE…
prabhdeepsingh/claude-plugins · sonu/skills/colors/SKILL.md · Color systems for web interfaces — OKLCH conversion and palette generation, contrast measurement (APCA/WCAG), gamut and…
prabhdeepsingh/claude-plugins · sonu/skills/debugging/SKILL.md · The scientific debugging loop — reproduce first, pull the real event from the repo's observability stack, one…
prabhdeepsingh/claude-plugins · sonu/skills/design-tree/SKILL.md · Make design decisions as an explicit branching tree — genuine alternatives, decisive rationale, rejected branches…
prabhdeepsingh/claude-plugins · sonu/skills/intent-interview/SKILL.md · Pre-spec intent extraction — one question at a time, each with a guess attached, until you can predict the user's…
ericrisco/rsc-harness · skills/performance/SKILL.md · Use when a page or interaction is slow for one user and the fix starts from a measurement — a failing Core Web Vital…
tomcounsell/ai · skills/valor/bama/skills/code/performance/SKILL.md · Make a macOS app fast by measuring hangs, stutter, CPU, memory, and energy, then moving work off the main thread and…
