Loading ToolBench

Reading the current corpus…

ig-trading-mcp MCP server · F · ToolBench

← Browse

ig-trading-mcp

github.com/kea0811/ig-trading-mcp

IG Trading API with MCP (Model Context Protocol) server for AI integration

<=2025-11-25STDIO20 tools

36/100

Definition Quality46

Protocol Readiness32

Supportability23

Spec posture

Inferred effective spec <=2025-11-25·current baseline 2026-07-28

Static source inference · medium confidence · detected: Logging

Deprecated protocol patterns detected

Summary

IG Trading MCP has 20 tools with significant definition gaps. Most tools lack input schemas (only 6 of 20 have visible schemas in the provided code). Descriptions are present but generic (avg ~40 chars, well below the 194-char baseline). Critical security issue: login/logout tools expose authentication as tool parameters rather than using server-side injection. No output schemas documented. Error handling is absent, no recovery guidance, no error categorization. Destructive tools (create_position, close_position) lack confirmation patterns. Tool composition is reasonable (verb_noun naming), but parameter descriptions are minimal. This server would not pass production code review.

Tools (20)

add_to_watchlistwriteauth50/100

Add market to watchlist

calculate_atrread only50/100

Calculate Average True Range indicator

calculate_supertrendread only50/100

Calculate Supertrend indicator for market data

close_positiondestructiveauth50/100

Close an open trading position

create_orderwriteauth50/100

Create a pending order

create_positiondestructiveauth50/100

Create a new trading position

delete_orderreversibleauth50/100

Delete a pending order

get_accountsread onlyauth40/100

Get list of accounts

get_activity

Issues (5)

CRITICALtool ↗

10 tools (get_accounts, get_positions, get_watchlists, get_working_orders, logout, login, and 4 others) have NO visible input schemas.

get_accountsget_positionsget_watchlistsget_working_ordersloginlogout

CRITICALsecret-injection ↗

Credentials (username, password, API key) exposed as tool parameters in login/logout. Per pattern:secret-injection, credentials must use server-side injection via environment variables or vault. Agent traces log all parameters, secrets leak into logs and prompt history.

loginlogout

CRITICALconfirmation-request ↗

Destructive tools (create_position, close_position, create_order, delete_order, update_order) lack confirmation patterns and dry-run support. No error guidance for recovery. Agents can execute irreversible trades without safeguards.

create_positionclose_positioncreate_orderdelete_orderupdate_order

Recommendations

Add complete input schemas for all 10 tools missing them (get_accounts, get_positions, get_watchlists, get_working_orders, login, logout, etc.). Include type, description, and constraints for every parameter.
Move credentials (username, password, API key) out of tool parameters. Use environment variables or a secure vault. Inject credentials server-side before calling the IG API. Document this in the tool description.
Add output schemas for all tools. Document what fields are returned (e.g., create_position returns dealId, direction, size, status). This enables LLMs to chain tools and extract required IDs.
Expand descriptions to 100-200 chars. Answer: What does it do? When should the LLM call it? What does it return? Example: 'Create a new trading position (BUY or SELL) on a specified market. Returns dealId for later reference. Use after search_markets() to find the epic.'
Add confirmation/dry-run pattern for destructive tools (create_position, close_position, create_order, delete_order, update_order). Implement a 'confirm' parameter or separate 'preview_*' tools so agents can review before executing irreversible trades.

Spec posture evidence

Inferred effective spec: <=2025-11-25.

Relies on Logging (deprecated) - log to stderr or use OpenTelemetry

Score history

Overall score trend

First recorded score · v2 rubric

36/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	36	<=2025-11-25	v2