Loading ToolBench
MCP server for deploying, querying, and operating HuaweiCloud infrastructure via the Model Context Protocol.
18 tools with significant definition gaps. Tool names follow verb_noun pattern (good), but descriptions are sparse and many lack depth. Input schemas present but parameter descriptions are minimal or in Chinese. No output schemas documented. Error handling absent. Security concerns: auth tool exposes credentials as parameters (AK/SK), openapi_request accepts arbitrary headers/bodies. No tool annotations (readOnlyHint, destructiveHint). Composition is reasonable (terraform tools chain well), but descriptions don't guide LLM selection or explain when to use each tool vs alternatives.
Authenticate with HuaweiCloud using AK/SK credentials
Query HuaweiCloud cost and billing information
Initialize a deployment workspace directory
Query the status of a deployment workspace
Retrieve a specific Terraform example
List available Terraform examples for HuaweiCloud resources
List existing HuaweiCloud resources in a region
Query and manage HuaweiCloud networking configuration
Credentials exposed as tool parameters: auth tool accepts 'ak' and 'sk' (HuaweiCloud Access Key and Secret Key) as string parameters. Credentials in tool params are logged, traced, and risk leaking into prompt history. Must use server-side secret injection via environment variables or secure vault.
No output schemas documented. Tools return results but LLMs cannot plan downstream calls without knowing field names and types. E.g., terraform_plan, terraform_apply, list_existing_resources lack documented return structures. Agents must guess what fields are available.
Parameter descriptions are minimal or in Chinese. 'deployment' param appears in 14 tools with description '部署工作区目录路径（绝对或相对）' (Chinese). English-only LLMs cannot parse Chinese descriptions. All descriptions must be in English and explain the parameter's role, format, and constraints.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 40 | 2026-07-28+ | v2 |
Make signed HTTP requests to HuaweiCloud APIs
Query Terraform provider schema for HuaweiCloud resources
Apply Terraform configuration to deploy infrastructure
Destroy Terraform-managed infrastructure
Initialize Terraform in the deployment workspace
Install Terraform in the deployment workspace
Generate a Terraform execution plan
Refresh Terraform state from HuaweiCloud
List resources in Terraform state
Show details of a resource in Terraform state
No error handling guidance. Tools do not document what errors can occur, whether they are retryable, or what the LLM should do next. E.g., terraform_apply can fail for many reasons (syntax error, permission denied, resource conflict) but no recovery guidance is provided.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). Terraform tools have clear risk profiles (terraform_apply is IRREVERSIBLE, terraform_destroy is DESTRUCTIVE) but are not annotated. LLMs cannot distinguish safe reads from dangerous writes without explicit hints.