The server defines 4 tools across prompts and project management. Tools have names following verb_noun conventions (prompt-get, prompts-list, project-switch, projects-list) and descriptions are present. However, schema quality is inconsistent: input schemas exist but lack depth in parameter descriptions, and output schemas are not documented. Descriptions are adequate but generic, missing LLM-optimization guidance on when to use each tool vs. alternatives. No tool annotations (readOnlyHint/destructiveHint) are declared despite risk taxonomy present. Error handling and recovery guidance are absent.

• Add output schema documentation to each tool. Include field names, types, and what each field represents. E.g., prompts-list should document: returns array of {id: string, name: string, description: string, ...}.
• Expand parameter descriptions with format hints and discovery guidance. E.g., prompt-get's 'id' parameter: 'The prompt identifier (string, 1-255 chars). Discover available IDs by calling prompts-list first.'
• Add error recovery hints to tool descriptions. E.g., prompt-get: 'If prompt not found, call prompts-list to verify the ID exists and matches exactly.'
• Declare tool annotations in schema: add 'readOnlyHint': true to prompts-list/projects-list; 'destructiveHint': true to project-switch if it modifies state.
• Enhance discovery tool descriptions with usage context. E.g., prompts-list: 'Call this first to discover available prompt templates, then use prompt-get to retrieve full details.'
• Document parameter constraints (min/max length, allowed characters, enums) in descriptions if not already in JSON Schema.
• Consider adding a 'prompt-describe' or 'project-describe' tool to return rich metadata without requiring a separate lookup, improving agent efficiency.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry