Loading ToolBench
HStorage CLI exposes 30 tools with basic schemas and descriptions, but falls short of production quality. Tool names lack consistent verb-noun structure (e.g., 'folder-share shares' instead of 'list_folder_shares'). Descriptions are present but generic (avg ~40 chars, well below the 194-char baseline). Parameter descriptions are sparse or missing entirely. No output schemas documented. Error handling is minimal, most tools return generic auth/validation errors without recovery guidance. The server is STDIO-only, which is a hard transport limitation. Overall, this reads as a CLI wrapper exposed via MCP rather than an agent-optimized tool suite.
Log in to HStorage
Log out from HStorage
Show auth status
Download a file from HStorage
Delete a file
Send file to an email
Get file information
List files
Tool names lack consistent verb-noun structure. Examples: 'folder-share shares' (should be 'list_folder_shares'), 'folder-share share' (should be 'create_folder_share'), 'team remove-member' (should be 'remove_team_member'). Inconsistent naming forces LLMs to reason about intent rather than parsing it from the name.
Descriptions are generic and under 50 characters for most tools (e.g., 'List files', 'Get file information'). Baseline is 194 chars. LLMs need context on WHEN to use each tool and what distinguishes it from similar tools. Current descriptions lack this guidance.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 40 | 2026-07-28+ | v2 |
Move a file to another folder
Update file state
Create folder
Delete folder
Get folder by UID
List folders
Update folder
Remove a folder share
Create a folder share
List folders shared with me
List folder shares
Update a folder share
Show team information
Invite a team member
Remove a team member
Show team storage status
Update team member storage allocation
Upload a file to HStorage
Delete current user account
Show user information
Get user settings
Update user settings
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields (e.g., after 'file list', what fields are returned? Is there pagination?). This forces agents to guess or make extra discovery calls.
Error handling is minimal. Tools return generic errors like 'AUTH_REQUIRED' or 'MISSING_CREDENTIALS' without recovery guidance. Per pattern:recovery-guide, errors should tell the LLM what to do next (e.g., 'Not logged in. Call auth login first.').
Destructive operations (file delete, folder delete, user delete, team remove-member) lack explicit confirmation or dry-run support. Per pattern:confirmation-request, irreversible operations should require explicit confirmation to prevent accidental data loss.