Communityコーディング&開発github.com

everyinc/ce-optimize

メトリクス駆動の反復最適化ループを実行します。測定可能な目標を定義し、並行実験を実行し、ハードゲートまたはLLM判定スコアに対して各結果を評価し、改善点を保持し、最適なソリューションに収束させます。クラスタリング品質、検索関連性、ビルドパフォーマンス、プロンプト品質など、体系的な実験が有効な測定可能な結果の最適化に使用します。

ce-optimize とは?

このスキルは、開発者やエンジニアがメトリクス駆動の反復最適化ループを実行するためのフレームワークを提供します。ユーザーはまず最適化したい測定可能な目標(例:クラスタリング品質、検索関連性、ビルドパフォーマンス、プロンプト品質)を定義します。次に、スキルは複数の並行実験を自動的に生成・実行し、各実験結果を事前定義されたハードゲート(合格/不合格のしきい値)やLLM-as-Judgeスコア(大規模言語モデルによる定性的評価)に対して測定します。改善が見られた結果のみを保持し、収束条件を満たすまでループを繰り返すことで、最適なソリューションへと導きます。主なプラットフォームとしてClaude Code、Cursor、CodexなどのAIエージェント環境で動作するよう設計されており、体系的な実験と評価の自動化により、人間の試行錯誤よりも効率的で再現性のある最適化を実現します。

対応~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/everyinc/compound-engineering-plugin/tree/main/skills/ce-optimize

Installed? Explore more コーディング&開発 skills: steipete/bluebubbles, steipete/eightctl, steipete/blucli · View all 6 →

お気に入りのAIに質問する

このエージェントスキルを事前に読み込んだ状態で新しいチャットを開きます。

ドキュメント

ce-optimize は何をしますか?

メトリクス駆動の反復最適化ループを実行します。測定可能な目標を定義し、並行実験を実行し、各結果をハードゲートまたはLLM-as-Judgeスコアに対して測定し、改善点を保持し、最適なソリューションに収束させます。クラスタリング品質、検索関連性、ビルドパフォーマンス、プロンプト品質など、体系的な実験が有効な測定可能な結果の最適化に使用します。

使用方法

  1. 目標の定義: 最適化したい測定可能なメトリクスを明確に定義します(例:クラスタリングのSilhouetteスコア、検索のNDCG@10、ビルド時間、プロンプト応答品質スコアなど)。
  2. 実験の設定: 並行して実行する実験の数と、各実験のパラメータ範囲を指定します。
  3. 評価基準の設定: ハードゲート(合格/不合格のしきい値)とLLM-as-Judgeスコアの評価基準を設定します。
  4. ループの実行: スキルが自動的に実験を生成・実行し、結果を評価し、最良のパラメータを保持します。
  5. 収束: 所定の収束条件(例:改善率が一定値を下回る、最大イテレーション数に達する)を満たすまでループを繰り返します。

対応プラットフォーム

  • Claude Code
  • Cursor
  • Codex
  • その他のAIエージェント環境

設定例

optimization:
  metric: "silhouette_score"
  experiments_per_iteration: 5
  hard_gate:
    min_silhouette: 0.3
  llm_judge:
    prompt: "クラスタリング結果の品質を1-10で評価してください"
  convergence:
    max_iterations: 10
    min_improvement: 0.01

注意事項

  • 実験のパラメータ範囲は適切に設定してください。範囲が広すぎると収束に時間がかかり、狭すぎると最適解を見逃す可能性があります。
  • LLM-as-Judgeを使用する場合、評価プロンプトは明確で一貫性のあるものにしてください。
  • ハードゲートは早期に不合格の実験を除外するために使用し、計算リソースを節約します。

ライセンス

MIT

Individual skills in this repo

This repo contains 20 individual skills — each has its own dedicated page.

everyinc/ce-agent-native-architecture

Build applications where agents are first-class citizens. Use this skill when designing autonomous agents, creating MCP tools, implementing self-modifying systems, or building apps where features are outcomes achieved by agents operating in a loop.

everyinc/ce-agent-native-audit

Run comprehensive agent-native architecture review with scored principles

everyinc/ce-brainstorm

Explore requirements and approaches through collaborative dialogue, then write a right-sized requirements document. Use when the user says "let's brainstorm", "what should we build", or "help me think through X", presents a vague or ambitious feature request, or seems unsure about scope or direction -- even without explicitly asking to brainstorm.

everyinc/ce-clean-gone-branches

Clean up local branches whose remote tracking branch is gone. Use when the user says "clean up branches", "delete gone branches", "prune local branches", "clean gone", or wants to remove stale local branches that no longer exist on the remote. Also handles removing associated worktrees for branches that have them.

everyinc/ce-code-review

Structured code review using tiered persona agents, confidence-gated findings, and a merge/dedup pipeline. In interactive mode it applies safe, verified fixes and commits them when the working tree is clean (it never pushes); in mode:agent it reports only and the caller applies. Use when reviewing code changes before creating a PR.

everyinc/ce-commit

Create a git commit with a clear, value-communicating message. Use when the user says "commit", "commit this", "save my changes", "create a commit", or wants to commit staged or unstaged work. Produces well-structured commit messages that follow repo conventions when they exist, and defaults to conventional commit format otherwise.

everyinc/ce-commit-push-pr

Commit, push, and open a PR with an adaptive, value-first description that scales in depth with the change. Use when the user says "commit and PR", "ship this", "create a PR", or "open a pull request". Also handles description-only flows ("write a PR description", "rewrite the PR body", "describe this PR") without committing or pushing.

everyinc/ce-compound

Document a recently solved problem to compound your team's knowledge or CONCEPTS.md, the project's shared domain vocabulary.

everyinc/ce-compound-refresh

Refresh stale learning and pattern docs under docs/solutions/ by reviewing them against the current codebase, then updating, consolidating, or deleting drifted ones. Use when the user asks to "refresh my learnings", "audit docs/solutions/", "clean up stale learnings", or "consolidate overlapping docs", or when ce-compound flags an older doc as superseded. Do not trigger for general refactor, debugging, or code-review work unless the user has explicitly pointed at docs/solutions/.

everyinc/ce-debug

Systematically find root causes and fix bugs. Use when debugging errors, investigating test failures, reproducing bugs from issue trackers (GitHub, Linear, Jira), or when stuck on a problem after failed fix attempts. Also use when the user says 'debug this', 'why is this failing', 'fix this bug', 'trace this error', or pastes stack traces, error messages, or issue references.

everyinc/ce-demo-reel

Capture a visual demo reel (GIF, terminal recording, screenshots) for PR descriptions. Use when shipping UI changes, CLI features, or any work with observable behavior that benefits from visual proof. Also use when asked to add a demo, record a GIF, screenshot a feature, show what changed visually, create a demo reel, capture evidence, add proof to a PR, or create a before/after comparison.

everyinc/ce-dhh-rails-style

This skill should be used when writing Ruby and Rails code in DHH's distinctive 37signals style. It applies when writing Ruby code, Rails applications, creating models, controllers, or any Ruby file. Triggers on Ruby/Rails code generation, refactoring requests, code review, or when the user mentions DHH, 37signals, Basecamp, HEY, or Campfire style. Embodies REST purity, fat models, thin controllers, Current attributes, Hotwire patterns, and the "clarity over cleverness" philosophy.

everyinc/ce-doc-review

Review requirements or plan documents using parallel persona agents that surface role-specific issues. Use when a requirements document or plan document exists and the user wants to improve it.

everyinc/ce-frontend-design

Build web interfaces with genuine design quality, not AI slop. Use for any frontend work - landing pages, web apps, dashboards, admin panels, components, interactive experiences. Activates for both greenfield builds and modifications to existing applications. Detects existing design systems and respects them. Covers composition, typography, color, motion, and copy. Verifies results via screenshots before declaring done.

everyinc/ce-gemini-imagegen

This skill should be used when generating and editing images using the Gemini API (Nano Banana Pro). It applies when creating images from text prompts, editing existing images, applying style transfers, generating logos with text, creating stickers, product mockups, or any image generation/manipulation task. Supports text-to-image, image editing, multi-turn refinement, and composition from multiple reference images.

everyinc/ce-ideate

Generate and critically evaluate grounded ideas about a topic. Use when asking what to improve, requesting idea generation, exploring surprising directions, or wanting the AI to proactively suggest strong options before brainstorming one in depth. Triggers on phrases like 'what should I improve', 'give me ideas', 'ideate on X', 'surprise me', 'what would you change', or any request for AI-generated suggestions rather than refining the user's own idea.

everyinc/ce-plan

Create structured plans for multi-step tasks -- software features, research workflows, events, study plans, or any goal that benefits from breakdown. Also deepens existing plans with interactive sub-agent review. Use when the user says 'plan this', 'create a plan', 'how should we build', 'break this down', or when a brainstorm doc is ready for planning. Use 'deepen the plan' or 'deepening pass' for the deepening flow. For exploratory requests, prefer ce-brainstorm first.

everyinc/ce-polish-beta

[BETA] Start the dev server, open the feature in a browser, and iterate on improvements together.

everyinc/ce-product-pulse

Generate a time-windowed pulse report on what users experienced and how the product performed - usage, quality, errors, signals worth investigating. Use when the user says 'run a pulse', 'show me the pulse', 'how are we doing', 'weekly recap', 'launch-day check', or passes a time window like '24h' or '7d'. Configures via .compound-engineering/config.local.yaml and saves reports to docs/pulse-reports/.

everyinc/ce-proof

Publish, view, comment on, and edit markdown via Proof (proofeditor.ai) — create a shareable doc, read a shared doc, and make comment/suggestion/block edits over its API. Use when the user says "view this in proof", "share to proof", "publish to proof", or wants a shareable markdown surface for a spec, plan, or draft, including publish handoffs from ce-brainstorm, ce-ideate, or ce-plan. Do not trigger on "proof" meaning evidence, math proofs, proof-of-concept, or "proofread this".

関連スキル