Agent skill

Concise Code

by octanejs in 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…

MITAuto-check passedDevelopment

Install Concise Code

skills CLI
$ npx skills add octanejs/octane --skill concise-code -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install octanejs/octane concise-code --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/octanejs/octane.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/concise-code .claude/skills/concise-code && rm -rf skills-src

Use ~/.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/

Facts

Skill name
concise-code
GitHub stars
1.5k
Token cost
~2.3k tokens
SKILL.md length
1,239 words
Files
1
Skills in repo
19
Repo updated
First seen
Licence
MIT

At a glance

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…

  • Works in 4 steps: Size the plan before writing it → Find the mechanism that already does it → Write in the existing idiom → …
  • Planning any change beyond a few lines
  • SKILL.md covers 1. Size the plan before…, 2. Find the mechanism that…, 3. Write in the existing idiom and 4. Simplify the diff before…, plus 1 more section
  • Calls git and pnpm

What it does

Concise Code is an agent skill from 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, and simplify the diff before handoff. Use when planning any change beyond a few lines, before readying a PR, or when reviewing agent-written code for bloat.

Its SKILL.md is about 2.3k 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

  • Planning any change beyond a few lines
  • Before readying a PR
  • Reviewing agent-written code for bloat

Example prompts

  • “/concise-code”

Workflow steps

4 steps, taken from the step headings in SKILL.md.

  1. Size the plan before writing it
  2. Find the mechanism that already does it
  3. Write in the existing idiom
  4. Simplify the diff before handoff

What it can do on your machine

Read from SKILL.md and the folder at commit 961638e. 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:

    • git
    • pnpm

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git 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

Concise Code loads about 2.3k tokens when it runs. Until then it costs about 92 tokens; SKILL.md has 1,239 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~92
When it runs · the whole SKILL.md, loaded when a task matches
~2.3k

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 961638e, republished under its MIT licence (© octanejs). 1,239 words, ~2,295 tokens.

Download SKILL.mdSave it as .claude/skills/concise-code/SKILL.md (or your agent's skills folder).
name
concise-code
description
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, and simplify the diff before handoff. Use when planning any change beyond a few lines, before readying a PR, or when reviewing agent-written code for bloat.

Skill: Concise code

Between July and October 2026 the whole client runtime grew from 37 kB to 181 kB gzip, and runtime.ts from about 17,000 to 52,000 lines. Most of it arrived one reasonable-looking change at a time: a branch per symptom, a second helper beside the first, a flag for a case the existing mechanism almost handled. No check fails on bytes, so the control is here: while you design, and before you hand off.

The goal is fewer concepts and less code to read, ship, and maintain, not fewer characters. Keep what earns its cost: fast paths that performance-audit asks for, comments that state invariants (the runtime's comments are its design spec), and tests of observable behavior. Code copied from an upstream React library follows octane-react-library-port, where fidelity to upstream wins over local brevity.

1. Size the plan before writing it

The first design that works is often the biggest. Before editing, write down what the plan adds: files, modules, exports, types, options, flags, Block or Scope fields, runtime imports, and lines of source by where they ship:

  • the client runtime (packages/octane/src outside compiler/) ships to every app;
  • the compiler and build plugins run in every build;
  • a binding ships to its users;
  • tests, scripts, and benchmarks ship nowhere, but are still read and kept.

When the plan adds a module, a mechanism, a public option, a hot-record field, or more than about 50 lines of runtime source, sketch at least two alternatives with their own estimates before writing any of them:

  • Extend the owner. Change the existing mechanism so it covers this case instead of adding a sibling that handles only this case. If the owner needs reshaping first, do that as a behavior-preserving commit, then make the now small change.
  • Remove the case. Change the data or the order of operations so the case cannot arise, rather than detecting and repairing it afterwards.
  • Fix the producer once. When several consumers would each need a guard, fix the one place that produces the bad value.
  • Pay only when used. Put feature-only code behind the capability or driver that owns it, so apps without the feature never ship it.
  • Delete. Check whether the change makes existing code, an older workaround, or a test-only path unnecessary.

Choose the smallest alternative that holds the contract and the hot-path rules. A larger one is fine when it is earned: name the mode it fixes, the hot path it keeps fast, or the test the smaller one fails. A fix much larger than its bug usually sits in the wrong layer or reimplements something, and a branch for one feature inside a general mechanism usually means the fix is too shallow. Either way, re-read the owner before writing it.

The thoughtWhat to do instead
"It is only a few more lines."Every change says that. Count what the whole plan adds.
"A new helper is cleaner than touching the old one."Two mechanisms for one job is the duplication. Change the old one.
"Safer to handle that case too."If the types or invariants exclude it, the branch is dead code.
"We will want this option later."Add it with its first caller.
"The review asked about this gap."A review told to find gaps always finds some. Close the ones that affect correctness or the contract.

2. Find the mechanism that already does it

Before writing a helper, git grep -n the DOM API, AST node type, error text, or concept it needs across packages/octane/src and the package you are changing. Read the nearest feature that solves a similar problem and follow its shape. These shared homes are the ones most often duplicated:

NeedUse
Which attributes render, their names and namespaces, void elementspackages/octane/src/dom-tables.js, shared by the compiler, client, and SSR
Delegated event namesevent-names.js
Unitless style props, style values, hyphenationstyle-values.js
URL attribute sanitizingsanitize-url.js
class and className compositionnormalizeClass in class-names.ts
Own-property checkshasOwnProp in has-own.ts
Posting a taskpostHostTask in host-task.ts; schedulePostPaint, createResizeObserver, and resumeOnSettle for their cases
Framework errors visible in productionthe catalog in packages/octane/error-codes/
Generated JavaScript@tsrx/core builders and clone helpers, per core-engineering
Mounting and flushing in core testspackages/octane/tests/_helpers.ts and the other _*.ts harnesses
Script path argumentsscripts/file-selection.mjs

If no shared home exists and a second place needs the same logic, move it to a module both import rather than copying it. A "keep in sync with" comment marks a copy; dom-tables.js replaced exactly such copies.

Show full SKILL.md (495 more words)Show less

3. Write in the existing idiom

  • One way per job. Match the surrounding naming, error construction, control flow, comment density, and test layout. A better pattern is its own refactor that converts the old sites, not a second dialect beside them.
  • Extract on a shared reason to change. Copies that merely look alike can stay; the third copy usually decides it. If you cannot name the helper plainly, it is not an abstraction yet. A helper whose boolean switches between two behaviors is two functions: the wrong abstraction costs more than the duplication it removed, so inline it back into its callers and extract again.
  • No speculative generality. No parameter, option, overload, or extension point without a caller in this change.
  • No defense against the impossible. Validate at boundaries such as the public API, parser input, the network, and user data, then trust internal callers. Do not catch only to rethrow or ignore, and do not add ?. or fallbacks for values that cannot be missing.
  • No pass-throughs. No wrapper, alias, or re-export that only renames.
  • No compensating state. Do not add a field or flag to work around an ordering problem; fix the order.
  • Comments say why. Do not restate the code, narrate the change, or record history; the commit message holds that.
  • Tests extend what exists. Add to the area's test file and fixture, use it.each or a table for cases that differ only in inputs, and never add a test-only branch to shipped code.

4. Simplify the diff before handoff

Read your diff as if you had to maintain it for someone else. Start with the numbers; git diff skips untracked files, so commit or git add -N them first:

bash
base=$(git merge-base HEAD origin/main)
git diff --stat "$base"
git diff --numstat "$base" | awk '$3 ~ "^packages/[^/]+/src/" { a += $1; d += $2 }
  $3 ~ "^packages/octane/src/" && $3 !~ "/compiler/" { ra += $1; rd += $2 }
  END { print "src +" a+0 " -" d+0 ", runtime +" ra+0 " -" rd+0 }'
git diff --diff-filter=A --name-only "$base"   # each new file needs a reason
git diff -U0 "$base" | grep -E '^\+.*\bexport\b'   # each new export needs a caller
git diff -U0 "$base" -- '*.ts' '*.tsx' '*.tsrx' '*.js' '*.mjs' | grep -E '^\+[^+]' |
  sed -E 's/^\+[[:space:]]*//' | awk 'length >= 40' | sort | uniq -cd | sort -rn | head -20

The last command lists added lines that occur more than once, which is usually pasted code. Then take each hunk in turn:

  1. Delete. Try removing each added branch, parameter, helper, field, and comment. Keep it only if the contract, a test, or a hot-path rule needs it.
  2. Reuse. git grep each new function's core call. If another function already makes it, call or extend that one.
  3. Collapse. Merge branches, functions, and tests that differ only by a value.
  4. Match. Compare each new construct with the nearest code doing the same job. A different style needs a reason.
  5. Proportion. Compare the diff with the size of the problem, and with the smallest alternative from step 1 now that the real diff exists.
  6. Bytes. For runtime changes, measure with the commands in performance-audit § Bundle bytes and justify any growth.

Stay inside the task: delete what your own change made unused, and report duplication outside your diff as a follow-up. Generated files belong to pnpm sync; never trim them by hand. When reviewing someone else's diff, give each finding a file:line, what is duplicated or unneeded, and the simpler form or the existing helper to call, and skip anything that would change behavior.

Report

Under Validation in the PR body:

md
- `concise-code`: src +A −D (runtime +R −S); new files and exports: <each, with its reason>; alternatives: <option, estimated size, why rejected>; bytes: <rows moved, or not applicable>

© 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/concise-code of octanejs/octane.

Open the folder on GitHubat commit 961638e

Compare with similar skills

Concise Code 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.

Concise Code compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Concise Code this skilloctanejs/octane1.5k—~2.3kAutomated safety check: PassMIT
Vercel Composition Patternssupabase/supabase111k58 repos~726Automated safety check: PassMIT
Finishing a Development Branchobra/superpowers297k5 repos~1.9kAutomated safety check: PassMIT
Typescript Advanced Typesrolling-scopes/rsschool-app10k25 repos~4.2kAutomated safety check: PassMPL-2.0
PR Babysitteropeninterpreter/openinterpreter69k3 repos~4.2kAutomated safety check: PassApache-2.0
Code Review ChecklistshareAI-lab/learn-claude-code78k5 repos~1.1kAutomated safety check: PassMIT

Similar skills

  • Official

    React composition patterns that scale. An agent skill from supabase/supabase.

    111k GitHub starsUsed in 58 repos~726 tokens
    DevelopmentAuto-check passed
  • 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.

    297k GitHub starsUsed in 5 repos~1.9k tokens
    DevelopmentAuto-check passed
  • 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.

    10k GitHub starsUsed in 25 repos~4.2k tokens
    DevelopmentAuto-check passed
  • 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.

    69k GitHub starsUsed in 3 repos~4.2k tokens
    DevelopmentAuto-check passed
  • 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.

    78k GitHub starsUsed in 5 repos~1.1k tokens
    DevelopmentAuto-check passed
  • 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.

    32k GitHub starsUsed in 4 repos~3.3k tokens
    DevelopmentAuto-check passed

More from octanejs/octane

All 19 skills in this repo
  • Performance Audit

    octanejs/octane

    Audit or defend Octane performance. An agent skill from octanejs/octane.

    1.5k GitHub stars~2.8k tokensUpdated today
    Auto-check passed
  • 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.

    1.5k GitHub stars~1.8k tokensUpdated today
    Auto-check passed
  • Handle Issue

    octanejs/octane

    Work a GitHub issue in the octane repo. An agent skill from octanejs/octane.

    1.5k GitHub stars~638 tokensUpdated today
    Auto-check passed
  • Migrate V8 To V9

    octanejs/octane

    Perform a complete @tanstack/react-table v8-to-v9 migration: hook and feature architecture, row-model slots, React state and subscriptions, rendering, composable tables, type helpers, and every…

    1.5k GitHub stars~4.7k tokensUpdated today
    Auto-check passed
  • Implement and verify new React-library ports or copied/rewritten React surfaces in Octane bindings from npm names or npm/GitHub links/lists.

    1.5k GitHub stars~2.5k tokensUpdated today
    Auto-check passed
  • Triage

    octanejs/octane

    Work out which part of the octane monorepo owns an unfamiliar failure or task.

    1.5k GitHub stars~374 tokensUpdated today
    Auto-check passed

Categories

Questions about Concise Code

What does Concise Code do?

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…. Concise Code is an agent skill from 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, and simplify the diff before handoff.

When should I use Concise Code?

Concise Code fits situations like: planning any change beyond a few lines; before readying a PR; reviewing agent-written code for bloat.

How do I install Concise Code in Claude Code?

Run `npx skills add octanejs/octane --skill concise-code -a claude-code`. Or copy the skill folder (.agents/skills/concise-code in octanejs/octane) into .claude/skills/concise-code in your project. Claude Code loads it when a task matches its description.

How do I install Concise Code in Codex?

Run `npx skills add octanejs/octane --skill concise-code -a codex`. Or copy the skill folder (.agents/skills/concise-code in octanejs/octane) into .agents/skills/concise-code in your project. Codex loads it when a task matches its description.

Can I use Concise Code 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 concise-code -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/concise-code, .gemini/skills/concise-code, .github/skills/concise-code and .opencode/skills/concise-code in your project.

What does Concise Code need to run?

Going by SKILL.md and its folder, Concise Code needs the command-line tools its instructions call (git and pnpm).

Does Concise Code access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Concise Code 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 Concise Code use?

Concise Code 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 Concise Code use?

About 2.3k tokens (SKILL.md is roughly 9.2k 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 Concise Code?

Skills that share tags, products or a category with Concise Code: 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 Concise Code?

octanejs (a GitHub organization) maintains it in octanejs/octane, which has 1,452 GitHub stars. The repository holds 19 skills in this directory. The repository was last updated on October 9, 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.