Linear Admin MCP has well-structured tool definitions with complete schemas and tool annotations present. All 5 tools have descriptions (10 - 200 chars), input schemas with typed properties, and proper annotation hints (readOnlyHint, destructiveHint, idempotentHint). However, parameter descriptions are minimal or absent, most params lack explanatory text beyond the name. Error handling is present but generic (no recovery guidance). Output schemas are not documented. Tool naming is clear and verb-based (linear_schema_find, linear_graphql_query, linear_project_setup_apply), but parameter names like 'contains' and 'variables' lack context. No pagination support visible. Composition is reasonable, tools are single-purpose, but chaining IDs are not explicitly documented in responses.

• Add detailed descriptions to all parameters. E.g., 'contains: Array of case-insensitive substrings to search for in Linear GraphQL type names and root field names. Repeat for multiple search terms. Example: ["User", "Team"].' This matches the pattern:tool-description baseline of 72 chars per param.
• Document output schemas for each tool. E.g., linear_schema_find returns {"types": [{"name": string, "description": string, "fields": [...]}], "rootFields": [...], "total": number}. Include field types and whether they are required.
• Enhance error messages with recovery guidance. E.g., 'GraphQL query failed: syntax error at line 3. Verify query is valid GraphQL. Call linear_schema_find to discover available types and fields.' This matches pattern:recovery-guide.
• Add limit and offset parameters to linear_schema_find. E.g., limit (1 - 100, default 20) and offset (default 0). Document that results are capped at 100 to prevent context window exhaustion.
• Implement dry-run mode for destructive tools. Add optional 'dry_run: boolean' parameter to linear_graphql_query and linear_project_setup_apply. When true, return the mutations that would be executed without applying them. This matches pattern:confirmation-request.