Vercel Composition Patterns
supabase/supabase
React composition patterns that scale. An agent skill from supabase/supabase.
Check that a diff, branch, or PR keeps Octane hot paths fast - stable V8 shapes and monomorphic sites, DOM work without forced layout, no new microtask hops or ad-hoc task posters - and that it…
Install Perf Review
$ npx skills add octanejs/octane --skill perf-review -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install octanejs/octane perf-review --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/octanejs/octane.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/perf-review .claude/skills/perf-review && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "perf-review" agent skill from https://github.com/octanejs/octane/tree/main/.agents/skills/perf-review into .claude/skills/perf-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "perf-review", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/octanejs/octane/tree/main/.agents/skills/perf-reviewType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add octanejs/octane --skill perf-review -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install octanejs/octane perf-review --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/octanejs/octane.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.agents/skills/perf-review .agents/skills/perf-review && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "perf-review" agent skill from https://github.com/octanejs/octane/tree/main/.agents/skills/perf-review into .agents/skills/perf-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "perf-review", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add octanejs/octane --skill perf-review -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install octanejs/octane perf-review --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/octanejs/octane.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.agents/skills/perf-review .cursor/skills/perf-review && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "perf-review" agent skill from https://github.com/octanejs/octane/tree/main/.agents/skills/perf-review into .cursor/skills/perf-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "perf-review", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/octanejs/octane.git --path .agents/skills/perf-review--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add octanejs/octane --skill perf-review -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install octanejs/octane perf-review --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/octanejs/octane.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.agents/skills/perf-review .gemini/skills/perf-review && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "perf-review" agent skill from https://github.com/octanejs/octane/tree/main/.agents/skills/perf-review into .gemini/skills/perf-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "perf-review", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install octanejs/octane perf-reviewInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add octanejs/octane --skill perf-review -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/octanejs/octane.git skills-src && mkdir -p .github/skills && cp -r skills-src/.agents/skills/perf-review .github/skills/perf-review && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "perf-review" agent skill from https://github.com/octanejs/octane/tree/main/.agents/skills/perf-review into .github/skills/perf-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "perf-review", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add octanejs/octane --skill perf-review -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install octanejs/octane perf-review --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/octanejs/octane.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.agents/skills/perf-review .opencode/skills/perf-review && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "perf-review" agent skill from https://github.com/octanejs/octane/tree/main/.agents/skills/perf-review into .opencode/skills/perf-review/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "perf-review", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
perf-review- GitHub stars
- 1.5k
- Token cost
- ~2.1k tokens
- SKILL.md length
- 986 words
- Files
- 1
- Skills in repo
- 19
- Repo updated
- First seen
- Licence
- MIT
At a glance
Check that a diff, branch, or PR keeps Octane hot paths fast - stable V8 shapes and monomorphic sites, DOM work without forced layout, no new microtask hops or ad-hoc task posters - and that it…
- Works in 5 steps: Choose the target → Classify the change before judging it → Read the mechanical candidates → …
- Development work in your project
- SKILL.md covers 1. Choose the target, 2. Classify the change before…, 3. Read the mechanical… and 4. Judge each dimension, plus 2 more sections
- Calls node, gh and pnpm
What it does
Perf Review is an agent skill from octanejs/octane. Check that a diff, branch, or PR keeps Octane hot paths fast - stable V8 shapes and monomorphic sites, DOM work without forced layout, no new microtask hops or ad-hoc task posters - and that it carries the evidence each risk needs. Use before readying a PR that touches packages/octane/src, compiler output, or binding hot paths, or when reviewing agent-written code.
Its SKILL.md is about 2.1k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
It sits in Development. The repository describes itself as: React’s programming model, compiled. The successor to Inferno. The licence is MIT.
When your agent uses it
- Development work in your project
Example prompts
- “/perf-review”
Workflow steps
5 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit efe12e3. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Shell commands in SKILL.md call:
nodeghpnpm
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md. Its commands use gh and pnpm, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Perf Review loads about 2.1k tokens when it runs. Until then it costs about 95 tokens; SKILL.md has 986 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from octanejs/octane at commit efe12e3, republished under its MIT licence (© octanejs). 986 words, ~2,123 tokens.
Download SKILL.mdSave it as
.claude/skills/perf-review/SKILL.md (or your agent's skills folder).- name
- perf-review
- description
- Check that a diff, branch, or PR keeps Octane hot paths fast - stable V8 shapes and monomorphic sites, DOM work without forced layout, no new microtask hops or ad-hoc task posters - and that it carries the evidence each risk needs. Use before readying a PR that touches packages/octane/src, compiler output, or binding hot paths, or when reviewing agent-written code.
Skill: Perf review
Use this to verify that a change applied the hot-path discipline in
performance-audit, which you load alongside this skill. Its references hold
the rules and the runtime code that already follows them:
V8 shapes,
DOM work, and
scheduling.
This is a validator. On someone else's PR, report findings and do not push fixes
unless asked. On your own diff before handoff, fix every must-fix finding and
run the review again.
1. Choose the target
node scripts/perf-review-scan.mjs # your worktree against its merge-base with origin/main
node scripts/perf-review-scan.mjs --head <branch> # a branch's committed diff
gh pr diff <number> --repo octanejs/octane | node scripts/perf-review-scan.mjs --diff -
node scripts/perf-review-scan.mjs packages/<binding>/src/ # a binding's hot path- The default scope is shipped runtime source under
packages/octane/src, excludingcompiler/. Path prefixes replace that scope; tests, fixtures, and benchmarks are always excluded.--jsonprints machine-readable findings. gh pr diffcarries only three lines of context, so the loop and read-after-write notes see less. A pasted diff also lacks the head'sruntime.ts, so--diffskipshot-class-shape. For a full review, check out the PR head in a worktree and use--head.
2. Classify the change before judging it
For each changed function, find its callers and decide how often it runs:
- Hot: per render, node, item, event, signal notification, or server
request. Framework fundamentals count as hot until the call graph shows
otherwise (
.rulesync/rules/core-engineering.md). - Cold: module initialization, once per root, error and abort paths, and development-only branches. A development-only branch must still not change a production shape.
Then note which dimensions apply: shapes and allocation, reachability, DOM, scheduling, and compiler output.
3. Read the mechanical candidates
The scan reports candidates on added lines with comments and strings blanked. Every candidate must end up as a finding or as a dismissal with a reason.
| Rule | Catches | Typical dismissal |
|---|---|---|
hot-class-shape | A BlockImpl, ScopeImpl, or LiteBlockImpl field that is a runtime class field, never assigned in the constructor, assigned conditionally, or assigned without a declaration | None. Holds on main; fix it |
hot-field-write | A write through a Block or Scope receiver, cast or not, to a field those classes do not declare | None. Holds on main; declare and initialize it |
delete-operator | delete on an object | Intentional dictionary or cold path |
conditional-shape | ...(cond ? {…} : {…}) or ...(cond && {…}) | Cold options object |
shape-mutation | Object.freeze, defineProperty, setPrototypeOf outside module-scope constants | Once per template or module, or a pinned exemplar |
holey-array | new Array(n) without .fill | Never indexed out of order and cold |
rest-or-arguments | A rest parameter or arguments | Cold branch, as in the HMR wrapper |
layout-read | Geometry reads, noting a DOM write earlier in the hunk | Batched measure phase, or a layout effect that reads before writing |
microtask-hop | queueMicrotask, Promise.resolve().then, noting an enclosing loop | One hop per burst that does no framework work per value |
await-as-yield | await of a settled value | Not used to yield |
schedule-render | A new scheduleRender call, noting an enclosing loop | Called once per burst, not per item or value |
animation-frame | requestAnimationFrame | Visual work meant to land before paint, with a timer fallback |
task-poster | MessageChannel, setTimeout(…, 0) or without a delay, setImmediate, postTask, requestIdleCallback | Extends an existing poster |
The scan cannot see the following, so check them by hand:
- Compiled output. Compile a representative fixture from
packages/octane/tests/_fixtures/before and after through the public compiler. Diff it for per-render closures, literals whose keys vary, extra runtime calls per node, changedbagNarity, and new runtime imports. - Reachability. A new reference from a hot or compiled path to a large
function or driver, a new import into
runtime.ts, or ahydratingguard that does not fold. - Polymorphism and representation. A hot function that now receives a new
receiver shape or argument type, returns a different shape, or stores a new
type in an existing field, such as a double in a Smi field or
undefinedin a numeric one. - Allocation. Closures, literals, spreads,
Array.from, or iterators created per item or per render. - DOM across functions. A read after a write that sits in a different function or hunk, and new per-element listeners or per-node DOM creation where a template clone would do.
- Scheduling semantics. An existing render request moved into a loop or a
producer, more frequent
drainPassivesBeforeRender, or a change to the contract indocs/differences-from-react.md§Scheduler.
Show full SKILL.md (322 more words)Show less
4. Judge each dimension
Answer these from the code. Cite file:line for every answer.
- Shapes: Is every new field on a hot record initialized at every allocation site, with identical keys and order across literal sites? Does any hot function gain a receiver map, argument type, or return shape?
- Allocation: What does the change allocate per render or per item? Can it be hoisted, reused, or replaced with an intrusive list?
- DOM: Are reads batched before writes, and outside the render walk? Is each
built subtree inserted once? Do resize callbacks that write go through
createResizeObserver? - Scheduling: Can any producer now render or commit per value or hop? Does a
new microtask chain do framework work at each step? Does a new task poster
duplicate
schedulePostPaint,actCheckpoint,createResizeObserver's poster, orresumeOnSettle, and does it survive hidden tabs andact()? Does the change alter the documented contract without a decision from #1864? - React divergences: Is anything you would flag a documented divergence in
docs/differences-from-react.md? Those are not defects. When a trade-off is ambiguous, prefer React semantics.
5. Require evidence
Each finding names the evidence it needs, from the table in performance-audit:
a one-map %HaveSameMap probe, allocation per call with pinned semi-space,
deterministic work counters, bundle rows from the CI report, the marker-task
commit count, or Event Timing in Chromium. Mark whether the PR provides it.
Run only the owning suite or a scratch probe locally, one at a time. Leave the
full pnpm test, benchmark sweeps, and browser suites to CI.
Report
## Perf review: <PR, branch, or worktree> (<base>..<head>)
Scan: `node scripts/perf-review-scan.mjs <args>` → <N> candidates.
Hot paths touched: <functions, with their frequency>.
| # | Severity | Location | Dimension | Finding | Required evidence | Provided? |
| --- | --- | --- | --- | --- | --- | --- |
| 1 | must-fix | packages/octane/src/runtime.ts:1234 | shapes | … | `%HaveSameMap` across modes | no |
Dismissed candidates:
- `packages/octane/src/runtime.ts:4567` [microtask-hop]: error-report path, once per uncaught error.
Not checked: <for example, browser latency, because no Chromium run>.- must-fix: breaks a rule on a hot path, or fails
hot-class-shapeorhot-field-write. - needs-evidence: may be fine, but the claim or the risk needs the listed measurement before the PR is ready.
- note: a cold-path observation or a follow-up.
A finding without a file:line and an argument for why the path is hot is not a
finding. When create-a-pr invokes this skill, paste the report into the PR
body's validation section.
© octanejs, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in .agents/skills/perf-review of octanejs/octane.
Open the folder on GitHubat commit efe12e3
Compare with similar skills
Perf Review next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Perf Review this skilloctanejs/octane | 1.5k | — | ~2.1k | Automated safety check: Pass | MIT | |
| Vercel Composition Patternssupabase/supabase | 111k | 58 repos | ~726 | Automated safety check: Pass | MIT | |
| Finishing a Development Branchobra/superpowers | 297k | 5 repos | ~1.9k | Automated safety check: Pass | MIT | |
| Typescript Advanced Typesrolling-scopes/rsschool-app | 10k | 25 repos | ~4.2k | Automated safety check: Pass | MPL-2.0 | |
| PR Babysitteropeninterpreter/openinterpreter | 69k | 3 repos | ~4.2k | Automated safety check: Pass | Apache-2.0 | |
| Code Review ChecklistshareAI-lab/learn-claude-code | 78k | 4 repos | ~1.1k | Automated safety check: Pass | MIT |
Similar skills
- Official
Finishing a Development Branch
obra/superpowers
Walks the last step of a branch: confirm tests pass, detect the git environment, ask how to integrate, carry out your choice and clean up the worktree.
Typescript Advanced Types
rolling-scopes/rsschool-app
Master TypeScript's advanced type system including generics, conditional types, mapped types, template literals, and utility types for building type-safe applications.
PR Babysitter
openinterpreter/openinterpreter
Watches an open GitHub pull request until it merges, handling review comments, diagnosing CI failures and retrying flaky checks along the way.
Code Review Checklist
shareAI-lab/learn-claude-code
Reviews code against a five-part checklist covering security, correctness, performance, maintainability and testing, and reports findings in a fixed format.
Greploop
onyx-dot-app/onyx
Iteratively improves a PR (GitHub), MR (GitLab), or shelved changelist (Perforce) until Greptile gives it a 5/5 confidence score with zero unresolved comments.
More from octanejs/octane
All 13 skills in this repoPerformance Audit
octanejs/octane
Audit or defend Octane performance. An agent skill from octanejs/octane.
Update Bindings
octanejs/octane
Audit one, several, or all existing Octane bindings; implement selected maintenance findings or remove redundant copied files with evidence matched to source ownership.
Concise Code
octanejs/octane
Keep a change small and in the existing idiom - size the plan and weigh smaller alternatives before writing it, reuse the mechanism that already owns the behavior instead of adding a parallel one…
Handle Issue
octanejs/octane
Work a GitHub issue in the octane repo. An agent skill from octanejs/octane.
Octane React Library Port
octanejs/octane
Implement and verify new React-library ports or copied/rewritten React surfaces in Octane bindings from npm names or npm/GitHub links/lists.
Triage
octanejs/octane
Work out which part of the octane monorepo owns an unfamiliar failure or task.
Related
Categories
Questions about Perf Review
What does Perf Review do?
Check that a diff, branch, or PR keeps Octane hot paths fast - stable V8 shapes and monomorphic sites, DOM work without forced layout, no new microtask hops or ad-hoc task posters - and that it…. Perf Review is an agent skill from octanejs/octane. Check that a diff, branch, or PR keeps Octane hot paths fast - stable V8 shapes and monomorphic sites, DOM work without forced layout, no new microtask hops or ad-hoc task posters - and that it carries the evidence each risk needs.
When should I use Perf Review?
Perf Review fits situations like: development work in your project.
How do I install Perf Review in Claude Code?
Run `npx skills add octanejs/octane --skill perf-review -a claude-code`. Or copy the skill folder (.agents/skills/perf-review in octanejs/octane) into .claude/skills/perf-review in your project. Claude Code loads it when a task matches its description.
How do I install Perf Review in Codex?
Run `npx skills add octanejs/octane --skill perf-review -a codex`. Or copy the skill folder (.agents/skills/perf-review in octanejs/octane) into .agents/skills/perf-review in your project. Codex loads it when a task matches its description.
Can I use Perf Review in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add octanejs/octane --skill perf-review -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/perf-review, .gemini/skills/perf-review, .github/skills/perf-review and .opencode/skills/perf-review in your project.
What does Perf Review need to run?
Going by SKILL.md and its folder, Perf Review needs the command-line tools its instructions call (node, gh and pnpm).
Does Perf Review access the network?
SKILL.md contains no URLs. Its commands use gh, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Is Perf Review safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Perf Review use?
Perf Review is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Perf Review use?
About 2.1k tokens (SKILL.md is roughly 8.5k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Perf Review?
Skills that share tags, products or a category with Perf Review: Vercel Composition Patterns (supabase/supabase, 111k stars), Finishing a Development Branch (obra/superpowers, 297k stars), Typescript Advanced Types (rolling-scopes/rsschool-app, 10k stars) and PR Babysitter (openinterpreter/openinterpreter, 69k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Perf Review?
octanejs (a GitHub organization) maintains it in octanejs/octane, which has 1,456 GitHub stars. The repository holds 19 skills in this directory. The repository was last updated on October 11, 2026.
Source: octanejs/octane on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.