Agent Quality Grading

Use when evaluating how well an agent completed real tasks, used tools, communicated, or produced assets; grade each dimension from traceable evidence and keep quality separate from reliability.

9 stars 3 forksAI, LLMs & Agents

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add sandbaseai/workbuddy-skill --skill agent-quality-grading
Manual install
git clone --depth 1 https://github.com/sandbaseai/workbuddy-skill.git /tmp/workbuddy-skill
mkdir -p ~/.claude/skills
cp -r /tmp/workbuddy-skill/skills/agent-quality-grading ~/.claude/skills/agent-quality-grading

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Agent Quality Grading

Use this skill to measure how well an agent served a user, independently of whether the underlying system was reliable. A polished message does not prove the task happened, and a successful tool call does not prove the result was useful. Grade only the evidence available for the requested scope.

## Establish scope and privacy

Define the time window, channels, agents, task classes, asset types, comparison baseline, and scoring audience. Obtain authorization for every trace and artifact. Prefer redacted or synthetic data; never copy credentials, personal data, private prompts, tokens, or raw secret-like content into the report. Record missing logs, clock assumptions, unavailable assets, and sampling limits.

Reconstruct each conversation or task from inbound request, agent turns, tool calls/results, generated artifacts, and final user-visible outcome in chronological order. Preserve IDs or hashes only when needed to join evidence. A claim such as “sent,” “updated,” or “deployed” requires a matching successful operation in the same trace; otherwise grade task completion as unverified or fabricated according to the repository policy.

## Grade four independent axes

Give each relevant task or channel a grade from A–F, with a short rationale and exact evidence reference:
Read the full skill on GitHub

Details

Skill name
agent-quality-grading
Path
skills/agent-quality-grading/SKILL.md
License
MIT
Repo updated
8 September 2026

FAQ

What does the Agent Quality Grading skill do?

Use when evaluating how well an agent completed real tasks, used tools, communicated, or produced assets; grade each dimension from traceable evidence and keep quality separate from reliability.

How do I install the Agent Quality Grading skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add sandbaseai/workbuddy-skill --skill agent-quality-grading. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Agent Quality Grading popular?

It lives in sandbaseai/workbuddy-skill, which has 9 GitHub stars and 3 forks.

Related skills

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.