Agent Evaluation
Evaluates agent performance across diverse tasks using a robust scoring system and detailed qualitative feedback.
Install on your platform
We auto-selected Claude Code based on this skill’s supported platforms.
Run in terminal (recommended)
claude mcp add glennguilloux-agent-evaluation npx -- -y @trustedskills/glennguilloux-agent-evaluation
Or manually add to ~/.claude/settings.json
{
"mcpServers": {
"glennguilloux-agent-evaluation": {
"command": "npx",
"args": [
"-y",
"@trustedskills/glennguilloux-agent-evaluation"
]
}
}
}Requires Claude Code (claude CLI). Run claude --version to verify your install.
About This Skill
What it does
The glennguilloux-agent-evaluation skill provides a framework for evaluating the performance of AI agents. It allows users to define evaluation criteria and assess an agent's responses against those criteria, providing structured feedback. This facilitates iterative improvement and ensures alignment with desired outcomes.
When to use it
- Testing new agent versions: Evaluate changes in agent behavior after updates or modifications.
- Comparing different agents: Objectively compare the performance of multiple AI agents on a specific task.
- Identifying areas for agent improvement: Pinpoint weaknesses in an agent's responses through structured assessment.
- Ensuring alignment with guidelines: Verify that an agent consistently adheres to defined rules or principles.
Key capabilities
- Definable evaluation criteria
- Structured response assessment
- Performance tracking over time
- Feedback generation for iterative improvement
Example prompts
- "Evaluate the following agent response: '[Agent Response]' against these criteria: [Criteria List]"
- "Compare the performance of Agent A and Agent B on this task: [Task Description] using the evaluation framework."
- "Assess this agent's response to 'User Query': '[User Query]' and provide feedback for improvement."
Tips & gotchas
The effectiveness of this skill relies on clearly defined and measurable evaluation criteria. Vague or subjective criteria will lead to inconsistent results.
Tags
TrustedSkills Verification
Unlike other registries that point to live repositories, TrustedSkills pins every skill to a verified commit hash. This protects you from malicious updates — what you install today is exactly what was reviewed and verified.
Security Audits
| Gen Agent Trust Hub | Pass |
| Socket | Pass |
| Snyk | Pass |
🌐 Community
Passed automated security scans.