agent-performance

To validate Claude Code agent YAML frontmatter, run the provided validation scripts to check for kebab-case names, valid descriptions, required fields, and valid model permissions. This process enforces validation steps to ensure safe, auditable agent configurations.

The best way to identify low-performing Ollama runtime agents is to analyze generated performance reports and compare different agent versions. This scoring process flags underperforming agents by evaluating their system prompts and generated outputs.

You generate actionable improvement recommendations for Claude Code agents by running performance analyses that evaluate body, frontmatter, template, and output scores. The skill outputs concrete steps to raise scores above required thresholds.

Yes, you can use this skill to compare agent versions and quantify scoring gaps. It analyzes performance reports across both Claude Code and Ollama agents to identify specific areas needing enhancement across the agent lifecycle.

Agent performance validation works for both Ollama runtime and Claude Code agents. The skill scopes validation across both platforms, evaluating YAML frontmatter, system prompts, and generated outputs to produce safe, auditable results.