doublegate

Bench Compare — Development skill for Claude Code

Development community

/bench-compare — statistical benchmark between two refs.

How to install Bench Compare

Installs to ~/.claude/skills/doublegate-master-claude-bench-compare/SKILL.md

Terminal

mkdir -p ~/.claude/skills/doublegate-master-claude-bench-compare && curl -fsSL https://raw.githubusercontent.com/doublegate/Master-Claude/HEAD/commands/bench-compare.md -o ~/.claude/skills/doublegate-master-claude-bench-compare/SKILL.md

Restart Claude Code, or start a new session, for it to be picked up.

What Bench Compare does

/bench-compare — statistical benchmark between two refs

Project-agnostic. Measures performance across two git refs and flags regressions/improvements. Derived from the recurring bench-compare command.

Usage

`/bench-compare [bench-target]`

Steps

Confirm a benchmark harness exists (master-core/lang/.md: cargo bench, pytest-benchmark, vitest bench, or a project script). If none, stop and say so.
For each ref: checkout (clean tree r

Alternatives in Development

Evo — turns your codebase into an autoresearch loop — discovers what to measure, instruments the benchmark, then run 1.4k ★
Aqe Benchmark 325 ★
Caveman Micro — 6-line caveman micro prompt (85 tokens) that outperformed the original 552-token skill 162 ★