jhelm provides 5 read-only Helm tools with complete input schemas and clear descriptions. All tools follow verb_noun naming (helm_template, helm_show_*, helm_lint). Descriptions are adequate (100-150 chars) and explain what each tool does. However, output schemas are not documented, LLMs cannot predict return structure. Parameter descriptions are present but minimal (10-30 chars); some lack format/constraint details. No error handling guidance. All tools are read-only (low risk), but lack the depth expected for production-grade tools.

• Document output schemas for all tools. E.g., helm_template returns {manifests: string, errors: string[], warnings: string[]}. helm_lint returns {errors: [{message, line}], warnings: [{message, line}]}.
• Expand parameter descriptions to include format constraints. E.g., namespace: 'Kubernetes namespace (1-63 lowercase alphanumeric + hyphens, must start/end with alphanumeric).'
• Add error recovery hints to tool descriptions. E.g., 'If chart path is invalid, verify it is absolute or relative to the current working directory. Common paths: ./my-chart, /opt/charts/my-chart, or a packaged .tgz file.'
• Document parameter relationships in helm_template. E.g., 'showOnly and skipTests are independent; includeCrds is ignored if the chart has no crds/ directory.'
• Add tool annotations (readOnlyHint: true, idempotentHint: true) to signal safety and enable agent optimizations.