Scoring was not performed

• Add complete input schemas (with types and required fields) to all 5 discovery tools (getViewer, getOrganization, getUsers, getLabels, getTeams) in package.json. Currently they have empty object schemas {}.
• Document input schemas for linear_searchIssues, linear_createIssue, linear_updateIssue, and linear_createComment. Include parameter types, descriptions, and required fields. These are critical WRITE operations.
• Consolidate tool definitions into a single registry or ensure all definitions in src/tools/handlers/*.ts are mirrored with explicit MCPToolDefinition exports like those in src/tools/definitions/*.ts.
• Add comprehensive error handling guidance to all tools, especially WRITE operations. Example: 'If you receive a 'duplicate issue' error, call linear_searchIssues() to check if this issue already exists.'
• Document output schemas for all tools. Based on cycle-tools.ts patterns, create explicit output_schema definitions for tools like linear_getIssues, linear_getIssueById, and read-only operations.
• Add enum constraints to 'state' and 'priority' parameters. Replace descriptive examples with formal JSON Schema enums: 'enum': ['planned', 'started', 'paused', 'completed', 'canceled'].
• Enforce pagination limits. For tools returning arrays (getIssues, getCycles, getProjectIssues), cap default limit at 25 and max at 100. Document in descriptions: 'Maximum: 100 issues per call (default: 25).'
• Enhance parameter descriptions for identifier fields. Specify format: 'issueId can be the Linear issue ID (e.g., 'PROJ-123') or internal UUID. System will resolve to internal ID automatically.'