skills.rest
alecs5am
alecs5amCOMMUNITY
Explore all skills by this author →
GitHub

workspace-eval

Score a project against its workspace rubric and generate a per-criterion report.

118|12|Updated May 5, 2026
Search Tags:#evaluation#scoring#rubric#append-only#ralphy#repair-loop#workspace-eval
One-click install
npx skills add https://github.com/alecs5am/ralphy --skill workspace-eval
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: workspace-eval
Source: https://github.com/alecs5am/ralphy/tree/main/.agents/skills/workspace-eval
Command: npx skills add https://github.com/alecs5am/ralphy --skill workspace-eval

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This surface provides a lightweight wrapper to run a workspace rubric evaluation for a project, producing a per-criterion scorecard and a pathway to repair when needed.

Core Features & Use Cases

  • Evaluates a project's alignment with its workspace rubric and exposes per-criterion scores.
  • Writes results to <project>/workspace-eval.json and <project>/workspace-eval-report.md using an append-only approach for auditability.
  • Integrates with the repair loop to hand off failing criteria for automatic remediation or human review.

Quick Start

Run ralphy workspace eval <project> to score it against its workspace rubric and produce the per-criterion report.

Frequently Asked Questions about workspace-eval

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

FAQPage Schema
How do I score a project against a workspace rubric?▼

To score a project against a workspace rubric, run ralphy workspace eval to generate a deterministic per-criterion scorecard. It evaluates alignment and writes results to workspace-eval.json and workspace-eval-report.md.

What is an append-only rubric evaluation scorecard?▼

An append-only rubric evaluation scorecard records per-criterion project alignment scores without overwriting history. This ensures auditability by persisting results exclusively to workspace-eval.json and workspace-eval-report.md files.

How do I hand off failing rubric criteria to a repair loop?▼

You can hand off failing rubric criteria to a repair loop by running a workspace evaluation to identify violations. The evaluation output feeds directly into automated remediation or human review pathways.

Can I use vision processing flags during workspace rubric scoring?▼

Yes, workspace rubric scoring respects configuration flags for vision processing and model selection. This allows deterministic evaluation tailored to specific multimodal project requirements.

Are workspace evaluation reports append-only?▼

Yes, workspace evaluation reports require append-only persistence for auditability. The evaluation writes scoring data continuously to workspace-eval.json and workspace-eval-report.md without modifying previous entries.

Related Skills

Explore More Software Engineering Skills→
yoshihiko555yoshihiko555

config-analyze

Parse Claude Code configuration files and generate a Japanese ANALYSIS.md with rubric evaluation.

Community
Intermediate
matteobortolazzomatteobortolazzo
1

babysit

Monitor GitHub pull requests for CI status and review feedback.

Community
Advanced
samuelmukotisamuelmukoti

llava

Integrate a vision encoder with a language model to generate multimodal responses from images.

Community
Advanced
miethemiethe

amcs-validator

Score composed artifacts against a blueprint rubric and output scores, issues, and pass.

Community
Advanced