eval — Claude Code command: what it does, how to install
Evaluates a plugin or skill directory using static checks and an LLM judge. Spots anti-patterns and scores quality up to 100 points.
/plugin marketplace add wshobson/agents
/plugin install plugin-eval@claude-code-workflowsWhat eval does
The command audits skill and plugin directories by running plugin-eval with uv. Passing --depth quick runs Layer 1 static analysis directly. It inspects directory layout and flags known anti-patterns immediately.
Standard depth adds a second layer through the eval-judge agent. The judge reads SKILL.md and reference documents to score trigger accuracy, orchestration fitness, and prompt clarity. Final scoring blends both layers across ten weighted dimensions. Without uv installed in your environment, static analysis will fail immediately.
The generated report lists dimension scores alongside detected anti-patterns. Scores of 60 earn a Bronze badge, while 90 or more unlocks Platinum. Anti-pattern penalties directly reduce the blended final total.
What it helps with
- Audit a skill directory for structural flaws and anti-patterns
- Benchmark prompt clarity and trigger precision using the
eval-judgeagent - Calculate an objective 100-point score for a Claude Code plugin
- Inspect
SKILL.mdand reference materials before distribution
Example prompts for the agent
- /eval ./skills/my-new-skill
- /eval ./plugins/linter --depth quick
How to install eval
1Plugin marketplace
/plugin marketplace add wshobson/agents/plugin install plugin-eval@claude-code-workflows2Manually: put the files in place
.claude/commands/eval.mdCall it with /eval. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/eval.mdCall it with /prompts:eval. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/eval.prompt.mdManually: put the files in place
.gemini/commands/eval.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about eval
How do I install eval in Claude Code?
Add the marketplace first with /plugin marketplace add wshobson/agents. Then install the plugin by running /plugin install plugin-eval@claude-code-workflows.
What does the eval command do?
It tests a skill or plugin directory across static code metrics and LLM evaluation. You receive a blended score out of 100, detected anti-patterns, and an earned quality badge.
What is the difference between quick and standard depth?
Quick depth runs Layer 1 static checks via uv run plugin-eval score and finishes instantly. Standard depth adds the eval-judge agent to evaluate SKILL.md and reference files, blending static scores with model judgement.
Source text
eval.md · MIT
---
description: Evaluate a plugin or skill for quality
argument-hint: <path> [--depth quick|standard]
---
Run the PluginEval quality evaluation on a plugin or skill directory.
## Usage
/eval <path> — evaluate at standard depth (static + LLM judge)
/eval <path> --depth quick — static analysis only (instant)
## Process
### Step 1: Run Static Analysis (Layer 1)
```bash
cd "${CLAUDE_PLUGIN_ROOT}"
uv run plugin-eval score {argument} --depth quick --output json
```
Parse the JSON output to get `composite.score`, `composite.dimensions`, and `layers[0].anti_patterns`.
### Step 2: LLM Judge (Layer 2) — if NOT --depth quick
Dispatch the `eval-judge` agent with the skill path:
> Evaluate the skill at: {resolved_path}
> Read the SKILL.md file and any references/ files, then score it on all 4 dimensions.
> Return your scores as JSON.
The judge returns scores for: triggering_accuracy, orchestration_fitness, output_quality, scope_calibration.
### Step 3: Compute Final Score
**If quick depth:** Report the Layer 1 results directly from the CLI output.
**If standard depth:** Blend Layer 1 and Layer 2 scores.
For each dimension, use these blend weights (Static:Judge):
- triggering_accuracy: 0.375:0.625
- orchestration_fitness: 0.125:0.875
- output_quality: 0.0:1.0 (judge only)
- scope_calibration: 0.353:0.647
- progressive_disclosure: 1.0:0.0 (static only)
- token_efficiency: 0.8:0.2
- robustness: 0.0:1.0 (judge only)
- structural_completeness: 0.9:0.1
- code_template_quality: 0.3:0.7
- ecosystem_coherence: 0.85:0.15
Dimension weights: triggering(0.25), orchestration(0.20), output(0.15), scope(0.12), disclosure(0.10), efficiency(0.06), robustness(0.05), structural(0.03), code_quality(0.02), coherence(0.02)
Final = sum(weight * blended_score) * 100 * anti_pattern_penalty
### Step 4: Present Results
```
## Overall Score: {score}/100 {badge}
## Layer Breakdown
| Layer | Score |
|-------|-------|
## Dimension Scores
| Dimension | Weight | Score | Grade |
|-----------|--------|-------|-------|
## Anti-Patterns Detected
## Recommendations
```
Badge thresholds: Platinum(90+), Gold(80+), Silver(70+), Bronze(60+)
© Seth Hobson · MIT · wshobson/agents
