Create and manage AI-powered trading bots. Use 'authenticate' first with your API key from vibetrader.markets/settings
VibeTrader MCP has a solid foundation with 9 tools that are mostly well-named and descriptive. All tools have descriptions (194-char baseline met), and all input parameters have type definitions and descriptions. However, there are critical gaps: (1) No output schemas are documented anywhere in the code, tools return JSON but what fields? what structure? (2) No error handling guidance, tools raise bare Exceptions without recovery hints (3) No parameter constraints (enums) despite tools like list_bots accepting a 'mode' parameter that should be constrained to 'paper'|'live'|None (4) Destructive tools (delete_bot) lack confirmation patterns or dry-run support (5) Tool composition could be improved, get_quote stands alone with no clear integration to bot creation/management. Tool naming is strong (all verb_noun), descriptions are adequate (10 - 200 chars each), but lack actionable guidance for LLM decision-making. Per-request auth via Bearer header is well-architected for multi-tenant safety, but this is not visible in tool definitions themselves.
Authenticate with your API key from vibetrader.markets/settings
Create a new trading bot using natural language
Delete a bot
Get details of a specific bot
Get portfolio positions and balance.
Get current quote for a stock
List all your trading bots.
No output schemas documented. Tools return JSON responses but the source code contains zero documentation of response structure, field types, or required fields. This forces LLMs to guess what data is available for downstream tool chaining.
Parameter constraints missing. 'list_bots' accepts mode='paper'|'live'|None but this is documented only in description text, not as an enum constraint. 'get_portfolio' mode parameter has same issue. Free-form strings invite hallucinated values.
Destructive tool lacks confirmation/dry-run. delete_bot can permanently destroy a trading bot with no confirmation step or dry-run option. Agents make mistakes, a single misunderstanding could delete real bots.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 60 | 2026-07-28+ | v2 |
Pause a running bot
Start a paused bot
Error handling does not guide recovery. Code raises bare Exceptions like 'Not authenticated!' or 'API Error (400): ...' without actionable next steps. LLM does not know: retry? user-fixable? fatal?
Tool composition incomplete. get_quote returns stock price but no integration path to create_bot. No batch tools (e.g. create_multiple_bots, pause_all_bots). Agents calling these sequentially waste tokens and latency.
Descriptions lack actionable guidance. 'Get portfolio positions and balance' tells what it does but not when to call it or what to do with the result. No dependency hints (e.g. 'Call this before creating a bot to check available capital').