Skip to content
mcp/skillhub

Agent Skill

Skill Judge skill

skill-judgesoftaworks/agent-toolkitskills/skill-judge/SKILL.md

Evaluate Agent Skill design quality against official specifications and best practices. Use when reviewing, auditing, or improving SKILL.md files and skill packages. Provides multi-dimensional scoring and actionable improvement suggestions.

2.6kMIT

always on

41 tokname + description

once used

6,485 tokSKILL.md body

bundled files

19 KB

repo stars

2.6kGitHub

Install the Skill Judge skill

Skills CLI (Claude Code, Codex, Cursor and others)

npx skills add https://github.com/softaworks/agent-toolkit --skill skill-judge

Claude Code, by hand (all projects)

git clone --depth 1 https://github.com/softaworks/agent-toolkit.git /tmp/agent-toolkit
mkdir -p ~/.claude/skills
cp -R /tmp/agent-toolkit/skills/skill-judge ~/.claude/skills/skill-judge

For one project only, copy the folder into .claude/skills/ in that project instead. The agent picks it up when a task matches its description.

What it costs your context

Agent Skills load in three steps. Tokens counted with the o200k tokenizer; your model's count will differ a little.

WhenWhat loadsCost
Every sessionName and description41 tok
When the agent uses itThe SKILL.md instructions6,485 tok
Only if opened1 bundled files (9 KB)varies

Until it is used, skill-judge costs about 33× less than the median MCP server we measured (1,363 tokens on every turn). Installing all 43 skills from softaworks/agent-toolkit would load 2,462 tokens of descriptions. Skills vs MCP study

Checks

  • The instructions are over 5,000 tokens, the Agent Skills specification's guideline for SKILL.md.
  • The instructions are over 500 lines; the specification suggests moving detail into reference files.

Checked 2026-10-22 · method

What's inside

The headings of its SKILL.md. Read the full file on GitHub.

  • Skill Judge
  • Core Philosophy
  • What is a Skill?
  • The Core Formula
  • Tool vs Skill
  • Three Types of Knowledge in Skills
  • Evaluation Dimensions (120 points total)
  • D1: Knowledge Delta (20 points) — THE CORE DIMENSION
  • D2: Mindset + Appropriate Procedures (15 points)
  • D3: Anti-Pattern Quality (15 points)
  • D4: Specification Compliance — Especially Description (15 points)
  • D5: Progressive Disclosure (15 points)
  • D6: Freedom Calibration (15 points)
  • D7: Pattern Recognition (10 points)
  • D8: Practical Usability (15 points)
  • NEVER Do When Evaluating
  • Evaluation Protocol
  • Step 1: First Pass — Knowledge Delta Scan
  • Step 2: Structure Analysis
  • Step 3: Score Each Dimension
  • Step 4: Calculate Total & Grade
  • Step 5: Generate Report
  • Common Failure Patterns
  • Pattern 1: The Tutorial

More skills from softaworks/agent-toolkit

All skills in this repo
  1. agent-md-refactorsoftaworks/agent-toolkit

    Refactor bloated AGENTS.md, CLAUDE.md, or similar agent instruction files to follow progressive disclosure principles. Splits monolithic files into organized, linked documentation.

    2.6k40 tok idle · 1,732 in use
  2. backend-to-frontend-handoff-docssoftaworks/agent-toolkit

    Create API handoff documentation for frontend developers. Use when backend work is complete and needs to be documented for frontend integration, or user says 'create handoff', 'document API', 'frontend handoff', or 'API documentation'.

    2.6k55 tok idle · 1,139 in use
  3. c4-architecturesoftaworks/agent-toolkit

    Generate architecture documentation using C4 model Mermaid diagrams. Use when asked to create architecture diagrams, document system architecture, visualize software structure, create C4 diagrams, or generate context/container/component/deployment diagrams. Triggers include "architecture diagram", "C4 diagram", "system context", "container diagram", "component diagram", "deployment diagram", "document architecture", "visualize architecture".

    2.6k82 tok idle · 2,505 in use
  4. codexsoftaworks/agent-toolkit

    Use when the user asks to run Codex CLI (codex exec, codex resume) or references OpenAI Codex for code analysis, refactoring, or automated editing. Uses GPT-5.2 by default for state-of-the-art software engineering.

    2.6k55 tok idle · 1,275 in use
  5. command-creatorsoftaworks/agent-toolkit

    This skill should be used when creating a Claude Code slash command. Use when users ask to "create a command", "make a slash command", "add a command", or want to document a workflow as a reusable command. Essential for creating optimized, agent-executable slash commands with proper structure and best practices.

    2.6k66 tok idle · 1,492 in use
  6. commit-worksoftaworks/agent-toolkit

    Create high-quality git commits: review/stage intended changes, split into logical commits, and write clear commit messages (including Conventional Commits). Use when the user asks to commit, craft a commit message, stage changes, or split work into multiple commits.

    2.6k54 tok idle · 522 in use
  7. crafting-effective-readmessoftaworks/agent-toolkit

    Use when writing or improving README files. Not all READMEs are the same — provides templates and guidance matched to your audience and project type.

    2.6k35 tok idle · 589 in use
  8. daily-meeting-updatesoftaworks/agent-toolkit

    Interactive daily standup/meeting update generator. Use when user says 'daily', 'standup', 'scrum update', 'status update', 'what did I do yesterday', 'prepare for meeting', 'morning update', or 'team sync'. Pulls activity from GitHub, Jira, and Claude Code session history. Conducts 4-question interview (yesterday, today, blockers, discussion topics) and generates formatted Markdown update.

    2.6k194 tok idle · 2,779 in use