Loading ToolBench
BM25-powered TeleBot Studio documentation search & REST API MCP server
Server has 26 tools with mostly complete schemas and descriptions. Naming follows verb_noun convention well (tbs_set_*, tbs_create_*, tbs_delete_*, etc.). Descriptions are present and substantive (avg ~150 chars), exceeding the 10-char minimum. However, several critical gaps reduce the score: (1) No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk classifications in metadata; (2) Output schemas are not documented, tools return JSON strings but structure is not formally specified; (3) Error handling lacks recovery guidance, most tools return generic error objects without actionable next steps; (4) Parameter descriptions are adequate but some lack format/constraint details (e.g., 'bot_id' described as 'numeric' but no validation rules shown); (5) Batch tools (tbs_batch_create_commands, tbs_batch_delete_commands) lack per-item failure reporting structure. Search tools (search_docs, search_examples, etc.) are well-described but top_k parameter lacks bounds (should specify 1-100 range). Agent-level tools (tbs_deploy_bot, tbs_setup_commands) are well-composed but orchestration logic is opaque to the LLM.
Retrieve an entire documentation page by name. Returns all sections and content for a specific page.
List all available documentation pages in the TeleBot Studio documentation.
Search the TeleBot Studio API reference documentation. Scoped search for API endpoints and methods.
Search the official TeleBot Studio documentation using BM25 ranking. Returns the most relevant documentation chunks matching your query.
Search for error messages and troubleshooting in the TeleBot Studio documentation. Scoped search for error handling and solutions.
Search for code examples in the TeleBot Studio documentation. Scoped search for practical usage examples.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk classifications in metadata. Tools marked DESTRUCTIVE (delete_bot, update_bot_token, delete_command, batch_delete_commands) should declare destructiveHint=true to guide LLM caution.
Output schemas not documented. Tools return JSON strings (e.g., tbs_deploy_bot, tbs_setup_commands, batch tools) but the structure of returned objects is not formally specified. LLMs cannot plan downstream calls without knowing response fields.
Error handling lacks recovery guidance. Error responses use generic _error_response() helper returning error code + message, but do not suggest next steps (e.g., 'Try search_users() first' or 'Retry with confirm=True'). Agents cannot self-correct.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 67 | <=2025-11-25 | v2 |
Search for function definitions in the TeleBot Studio documentation. Scoped search for specific functions.
Search the TeleBot Studio library documentation. Scoped search for library functions and classes.
Create multiple commands on a bot in sequence. Each command is created independently. If one fails, the rest continue. A summary with per-command results is returned.
Delete multiple commands from a bot in sequence. This is a destructive operation. Each command is deleted independently. If one fails, the rest continue. A summary is returned.
Create a new TeleBot Studio bot with a Telegram bot token. After creation, the bot ID is automatically set as the active bot for the current session. Get your bot token from @BotFather on Telegram.
Create a new command on a TeleBot Studio bot.
Check the current session credential status. Returns whether the API key and Bot ID are set (key is masked). Use this to verify your credentials before performing API operations.
Soft-delete a TeleBot Studio bot (marks as deleted, preserves data). This is a destructive operation. Set confirm=True to execute. When confirm=False (default), returns a preview of what will happen.
Delete a command from a TeleBot Studio bot. This is a destructive operation. Set confirm=True to execute. When confirm=False, returns a preview.
Complete bot deployment: create bot, add commands, and start it. This is an agent-level tool that orchestrates multiple API calls in sequence. The AI uses this to fulfill high-level requests like "build me a bot" or "deploy a premium system." The bot ID from creation is automatically used for subsequent steps.
Get details of a specific command by name.
List all commands for a TeleBot Studio bot.
Restart a TeleBot Studio bot (stop + start).
Set your TeleBot Studio API key for the current session. The API key is stored in memory only — never persisted to disk, never logged in cleartext. It is lost when the server restarts. You can find your API key at: telebotstudio.com → Account Settings → API
Set the active TeleBot Studio Bot ID for the current session. This bot ID will be used for all bot-scoped operations (commands, start, stop, restart) unless overridden explicitly.
Bulk create commands on an existing bot. This is an agent-level tool for setting up multiple commands at once.
Start a TeleBot Studio bot (set webhook).
Stop a TeleBot Studio bot (remove webhook).
Update the Telegram bot token for an existing bot. The bot is automatically restarted after the token update. This is a destructive operation. Set confirm=True to execute.
Update an existing command's code. This is a destructive operation. Set confirm=True to execute. When confirm=False, returns a preview.
Search tools (search_docs, search_examples, search_api, search_library, search_functions, search_errors) lack bounds on top_k parameter. Should specify 1-100 range to prevent LLM from requesting 1000+ results and exhausting context.
Batch tools lack per-item failure reporting. tbs_batch_create_commands and tbs_batch_delete_commands return a summary but structure is not documented. If 1 of 50 commands fails, LLM cannot determine which one or retry selectively.