Server has 33 tools with basic descriptions and parameter schemas. Naming follows verb_noun convention (list_, get_, setup_, clear_). However, critical gaps exist: (1) Output schemas are completely undocumented, no tool specifies what fields are returned, forcing LLMs to guess structure; (2) Parameter descriptions are minimal (10-30 chars) and lack constraints, ranges, or format guidance; (3) No error handling guidance, tools return raw strings without recovery hints; (4) Sensitive credential handling in setup_credentials lacks security warnings; (5) Many tools lack pagination despite returning lists. Per-tool analysis: setup_credentials, get_credentials_status, clear_credentials have adequate descriptions but no output schema. All 30 resource/compute/network/storage tools follow consistent naming but descriptions are generic ('List all X', 'Get detailed X configuration') without context on when to use each vs. similar tools. No tool documents whether results are paginated, what fields are included, or how to chain results to downstream tools.
Clear stored credentials from session
Get detailed configuration of a specific static IP address
Get detailed information about a specific cloud
Check current credentials status
Get detailed disk configuration
Get detailed information about a specific disk type
Get detailed information about a specific folder
No output schemas documented for any tool. LLMs cannot infer what fields are returned, forcing guesswork on downstream tool chaining. All 33 tools lack return type documentation.
Parameter descriptions are generic and lack constraints. E.g., 'folder_id' described as '20 character alphanumeric string' but no validation guidance. No min/max for numeric params, no enum constraints for choice params.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 43 | <=2025-11-25 | v2 |
Get detailed configuration of a specific gateway
Get detailed configuration of a specific image
Get detailed configuration of a specific network
Get detailed configuration of a specific route table
Get detailed security group configuration
Get detailed snapshot configuration
Get detailed configuration of a specific subnet
Get detailed VM configuration
Get organization context with available clouds and folders hierarchy
Get detailed information about a specific zone
List all availability zones
List all disk types in specified zone or all zones
List all images in specified folder
List all disk snapshots
List all storage disks
List all virtual machines
List all static IP addresses in specified folder
List all gateways in specified folder
List all VPC networks
List all route tables in specified folder
List all security groups
List all VPC subnets
List all clouds in organization or all accessible clouds
List all folders in specified cloud or all accessible folders
Setup Yandex Cloud credentials for the session
Suggest appropriate scope (folder/cloud/organization) for resource queries
No pagination support documented. All list_* tools accept optional folder_id but do not specify limit, offset, page_size, or next_cursor parameters. Large result sets will blow context windows.
No error handling guidance. Tools return raw strings (e.g., 'Invalid IAM token format') but do not guide LLM on recovery steps. No distinction between retryable, user-fixable, and fatal errors.
Naming inconsistency: 'get_cloud_details_info' and 'get_folder_details_info' use redundant '_info' suffix. Should be 'get_cloud_details' and 'get_folder_details' to match other get_* tools.