LeadGrowGTM

Benchmarking Loop — Testing skill for Claude Code

Testing community

Run a benchmark spec through the benchmarking loop (sweep or climb) and write a ranked variant ledger.

How to install Benchmarking Loop

Installs to ~/.claude/commands/leadgrowgtm-loop-engineer-benchmarking-loop.md

Terminal

mkdir -p ~/.claude/commands && curl -fsSL https://raw.githubusercontent.com/LeadGrowGTM/loop-engineer/HEAD/.claude/commands/benchmarking-loop.md -o ~/.claude/commands/leadgrowgtm-loop-engineer-benchmarking-loop.md

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

What Benchmarking Loop does

description: Run a benchmark spec through the benchmarking loop (sweep or climb) and write a ranked variant ledger argument-hint: "[goal text | | --resume ]"

/benchmarking-loop

Thin front-door **router** for the benchmarking loop - the harness's second goal path, slotted beside `/write-goal-prompt` (ADR-0004). It **runs** a frozen benchmark spec; it does NOT re-author one. Authoring lives in the shared grill + the lazy `skills/write-goal-prompt/references/benc

Alternatives in Testing

Claude Code Spec Workflow — Automated Kiro-style Spec workflow for Claude Code 3.6k ★
Spec Driven Development — Drive Kandev feature work through spec, plan, independent tasks, implementation, QA, and verification 697 ★
Al Spec.Create.Prompt — Create or revise the canonical AL technical specification through AL Spec Agent, preserving approved architect 109 ★