Loading ToolBench
MCP server for Framedeck — connect any AI assistant to your content production pipeline
framedeck-mcp provides 20 tools with consistent naming (verb_noun pattern) and basic descriptions. All tools have input schemas with type definitions. However, descriptions are generic and lack LLM-optimization guidance. Parameter descriptions are minimal (often just restating the name). Output schemas are undocumented. Error handling is absent, no recovery guidance, no actionable error messages. Missing tool annotations (readOnlyHint, destructiveHint). The server accepts both IDs and names (good), but lacks validation constraints, enums, and format specifications. No pagination support despite list tools. Composition is reasonable (single responsibility per tool), but chaining support is weak, responses don't guarantee all IDs needed for downstream calls.
Create a new board (project or production)
Create a new card in a column
Create a new column (stage) in a board
Create a new label (tag)
Delete a board and all its contents
Delete a card
Delete a column and move its cards to another column
No output schemas documented. LLMs cannot infer what fields list_boards, get_card, or create_column return, forcing them to guess downstream field names and risk broken tool chains.
Descriptions are generic and under 100 chars. E.g., 'List all boards (projects or productions) for the authenticated user' lacks guidance on when to call this vs get_board, what structure is returned, or pagination behavior.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). LLMs cannot distinguish safe reads from destructive deletes without explicit hints. delete_board, delete_column, delete_card, delete_label should be marked destructive.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 54 | 2026-07-28+ | v2 |
Delete a label
Get details of a specific board by name or ID
Get details of a specific card by name or ID
List all boards (projects or productions) for the authenticated user
List all cards in a column or board
List all columns (stages) in a board
List all labels (tags) available to the user
Move a card to a different column or position
Set the active board context for subsequent operations
Update board properties (name, icon, color, etc.)
Update card properties (title, description, labels, assignees, etc.)
Update column properties (name, color, etc.)
Update label properties (name, color, etc.)
No error handling or recovery guidance. If a tool fails (e.g., board not found, invalid color hex), the LLM receives no actionable error message or next steps. Errors should guide the agent to search_* tools or suggest corrections.
List tools lack pagination (limit, offset, page_size, next_cursor). Returning all boards/columns/cards/labels risks context window exhaustion. Descriptions should state result limits and pagination support.