Server has 9 tools with explicit schemas and descriptions. However, most descriptions are generic or task-focused rather than LLM-optimized. The run-gcp-code tool has an extremely long description (1000+ chars) that embeds implementation rules rather than explaining when/why to use it. Parameter descriptions are present but often lack constraints (enums, ranges, formats). No output schemas are documented. Error handling is minimal, tools return raw API responses without recovery guidance. Naming is verb-forward (list-, get-, run-) which is good, but descriptions don't explain tool selection criteria or dependencies.

• Rewrite run-gcp-code description to 50-150 chars: 'Execute arbitrary GCP code in TypeScript using official client libraries. Use when other tools don't cover your query.' Move implementation rules to a separate prompt or documentation.
• Add output schema documentation for all tools. Example for list-projects: 'Returns {projects: [{id: string, name: string, status: string}], total: number}'.
• Add numeric constraints to parameters: get-cost-forecast months should be 1-36; get-logs pageSize should be 1-1000.
• Add error recovery guidance to descriptions. E.g., get-billing-info: 'Returns billing account info. If not found, ensure billing is linked to the project via GCP Console.'
• Clarify tool selection in descriptions. E.g., get-billing-info returns current charges; get-cost-forecast predicts future spend; get-billing-budget shows budget thresholds.