marino39

Workflow Eval — Code Quality skill for Claude Code

Code Quality community

Evaluate the workflow — deterministic lint over the instruction files plus an optional scored live run of /new-task against frozen fixture tasks, with ablation A/B. Expensive; opt-in, not auto-schedul.

How to install Workflow Eval

Installs to ~/.claude/skills/marino39-software-developer-workflows-workflow-eval/SKILL.md

Terminal

mkdir -p ~/.claude/skills/marino39-software-developer-workflows-workflow-eval && curl -fsSL https://raw.githubusercontent.com/marino39/software-developer-workflows/HEAD/commands/workflow-eval.md -o ~/.claude/skills/marino39-software-developer-workflows-workflow-eval/SKILL.md

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

What Workflow Eval does

description: Evaluate the workflow — deterministic lint over the instruction files plus an optional scored live run of /new-task against frozen fixture tasks, with ablation A/B. Expensive; opt-in, not auto-scheduled.

Workflow Eval: $ARGUMENTS

Measure the workflow so complexity decisions are evidence-based (the "measure & iterate" discipline). Two layers: a cheap deterministic **lint**, and an expensive **live outcome-eval**. `$ARGUMENTS` flags:

--lint-only — run only Layer 1 and

Alternatives in Code Quality

Brooks Sweep — Run all Brooks-Lint analyses in one pass, auto-fix safe findings, confirm risky ones 1.5k ★
Claude Command Suite — Professional slash commands for Claude Code that provide structured workflows for software development tasks 1.1k ★
Lint Task — Lint coder-eval Task YAML 166 ★