alfredoperez

Bench Capture — Development skill for Claude Code

Development community

Score the 2 bench variant folders for one size (build/acceptance/regression/conventions/capture + rubric + cross-solution review), record, report, and reset.

How to install Bench Capture

Installs to ~/.claude/commands/alfredoperez-speckit-companion-bench-capture.md

Terminal

mkdir -p ~/.claude/commands && curl -fsSL https://raw.githubusercontent.com/alfredoperez/speckit-companion/HEAD/.claude/commands/bench-capture.md -o ~/.claude/commands/alfredoperez-speckit-companion-bench-capture.md

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

What Bench Capture does

allowed-tools: Bash(node examples/todo-claude/bench/run-all.mjs:*), Agent, Workflow, AskUserQuestion description: Score the 2 bench variant folders for one size (build/acceptance/regression/conventions/capture + rubric + cross-solution review), record, report, and reset

Your task

After you've run a size through the two folders in VS Code, capture the evals, append to history, regenerate the report, and reset the folders for the next round.

1. Resolve the size

From `$ARGUMENTS`

Alternatives in Development

Extraction Judge.Prompt — Extraction Judge — Fact Durability Rubric 189 ★
Bench — Run performance benchmarks for RLM-Claude-Code 86 ★
Batch Benchmark — Launch the 4-variant batch on the currently-free local GPUs (parallel when N>=4, otherwise sequential waves of 68 ★