Loading ToolBench
Read-only MCP server for the Scalable Capital broker, via the official `sc` CLI.
Server has 10 tools with basic descriptions and minimal input schemas. Most tools are read-only portfolio queries with empty input schemas ({}), which is appropriate for their use case. However, descriptions are terse (10-60 chars), parameter descriptions are sparse, and output schemas are completely undocumented. The two parameterized tools (search_security, security_news) have better schemas with Zod validation, but lack output documentation. No error recovery guidance, no tool annotations, and no composition hints. Typical of a STDIO-only broker integration, functional but not production-grade.
Portfolio analytics and performance breakdown.
Check whether the Scalable CLI is installed and logged in. Run this first if any other tool fails.
Current holdings with quantities and market values.
Portfolio overview: total value, profit/loss, and asset allocation.
Your active price alerts (read-only).
Your configured savings plans (read-only).
Output schemas completely undocumented. No tool returns a documented schema, LLMs cannot predict response structure or plan downstream calls. All 10 tools return raw CLI JSON output without field documentation.
Descriptions are too terse (10-60 chars). Rubric baseline is 194 chars average; these descriptions lack context for when/why to call each tool. E.g., 'Portfolio overview: total value, profit/loss, and asset allocation.' (60 chars) does not explain when to use it vs analytics or holdings.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 56 | 2026-07-28+ | v2 |
Search for a tradable instrument by name or ticker.
Recent news for a specific instrument, identified by ISIN.
Recent broker transactions.
Instruments on your watchlist (read-only).
No error recovery guidance. readCommand() returns isError=true with raw CLI error messages, but does not guide the LLM on next steps. E.g., 'CLI not found' should suggest 'Run check_setup first' or 'Install Scalable CLI'.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). All tools are read-only, but this is not declared in the tool definition, LLMs cannot infer safety properties.
No pagination or result limits documented. Tools like holdings, transactions, watchlist may return large result sets that blow context windows. No limit parameter, no cursor, no total count in response.