skills.rest
xotong
xotongCOMMUNITY
Explore all skills by this author →
GitHub

ce-optimize

Automate metric-driven iterative optimization with parallel experiments and gating rules.

2|Updated May 8, 2026
Search Tags:#optimize#code#gate#git-worktree#experiments#judge#measurement-harness
One-click install
npx skills add https://github.com/xotong/claude-marketplace --skill ce-optimize-xotong
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ce-optimize
Source: https://github.com/xotong/claude-marketplace/tree/main/plugins/compound-engineering/skills/ce-optimize
Command: npx skills add https://github.com/xotong/claude-marketplace --skill ce-optimize-xotong

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires python3, git, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Metric-driven experimentation to optimize hard system metrics and qualitative outputs using gates and LLM judges, with durable artifacts and crash-safe logging.

Core Features & Use Cases

  • Phase-based optimization with baseline measurement, hypothesis backlog, and batch evaluation for quick iteration.
  • Supports both hard (objective) metrics and LLM-based judge metrics for qualitative targets.
  • Deterministic isolation via Git worktrees and explicit mutable/immutable scope, plus crash-safe experiment logging.

Quick Start

Define a ce-optimize spec and run the baseline measurement to bootstrap the optimization loop.

Frequently Asked Questions about ce-optimize

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate metric-driven code optimization with parallel experiments?▼

Metric-driven code optimization is automated by running parallel experiments across Git worktrees, evaluating outcomes with gating rules, and consolidating results into a durable experiment log for iterative improvement.

Can I use an LLM judge for qualitative code optimization instead of hard metrics?▼

Yes, an LLM judge can evaluate qualitative code optimization targets. The system supports both hard objective metrics and LLM-based judge scores, applying stratified sampling and gating rules to assess qualitative outcomes.

How does Git worktree isolation work for running optimization experiments?▼

Git worktree isolation provides deterministic environments for optimization experiments by explicitly separating mutable and immutable scope, ensuring parallel tests run without interfering with your primary working directory.

Do I need Python and Git to run automated optimization loops?▼

Yes, Python3 and Git are required dependencies. The optimization loop coordinates spec generation, hypothesis backlogs, and worktree-backed experiments using Python scripts to execute and log iterative tests safely.

What is the best way to track and audit iterative code optimization results?▼

The best way to track iterative code optimization is through a crash-safe experiment log. It coordinates spec generation, hypothesis backlogs, and batch evaluations, creating durable artifacts for an auditable optimization loop.

Related Skills

Explore More Software Engineering Skills→
yu13130122297yu13130122297

scout

Identify uncertainties in research framing, datasets, metrics, and baselines.

Community
Intermediate
riethmayerriethmayer
2

skill-creator

Create, evaluate, and refine Claude Skills through iterative SKILL.md drafting and eval loops.

Community
Advanced
himanshu231204himanshu231204
2

agent-eval

Benchmark coding agents with reproducible metrics and git worktree isolation.

Community
Intermediate
applied-artificial-intelligenceapplied-artificial-intelligence
81

llm-evaluation

Evaluate LLM applications with automated metrics and LLM-as-judge scoring.

Official
Advanced