Loading ToolBench
Research-grounded metacognitive reasoning tools for AI agents. Structured step-by-step thinking with steel-manning verification via MCP. Combines sequential decomposition with cognitive science frameworks. Backed by 43+ papers.
Steelmind MCP provides two metacognitive tools with complete input schemas and detailed descriptions. Both tools are read-only and follow a clear naming convention (verb_noun). However, output schemas are not documented, and the tools lack error handling guidance. Descriptions are well-written (200+ chars) and explain WHEN to use each tool, but there is no guidance on what the LLM should do if a tool call fails or produces unexpected output. The server is well-structured and uses proper TypeScript with ESLint, but the tool definitions themselves are minimal, they log reasoning steps without returning structured data that downstream tools could consume.
Use this tool to record a structured reasoning step. It will not obtain new information or change any state — it appends your thought to the log. Use it when you need to: process results from previous tool calls before acting, plan your approach to a multi-step task, analyze a complex situation before deciding, or navigate environments with detailed policies. Do NOT use for simple single-step tasks or restating without analysis. When thinking, ask yourself: What am I assuming? What evidence supports this? What's my plan, and what could go wrong? You can adjust totalThoughts up or down as your understanding deepens. When you set nextThoughtNeeded to false, use the verify tool to challenge your conclusion before acting.
Use this tool to challenge and evaluate your reasoning before committing to an action. It will not obtain new information or change any state — it logs your critical self-assessment. Use it when you need to: check if your planned action complies with all requirements, validate reasoning before committing, assess edge cases, or evaluate tool results for correctness. Do NOT use to confirm what you are already confident about. When verifying, steel-man the opposition: What is the strongest argument that your conclusion is wrong? If you can't defeat it, reconsider. If your verification reveals a flaw, use the think tool to revise your approach.
Output schemas not documented. Tools return text responses but LLMs cannot infer structure for downstream composition.
No error handling guidance. If a tool call fails or produces invalid input, descriptions do not tell the LLM what to do next.
Parameter descriptions lack format/constraint details. 'thought' and 'concern' accept free-form strings with no guidance on length, structure, or validation rules.
No tool composition support. Tools are read-only logging utilities; they do not return IDs or references that other tools could consume.
Missing tool annotations. Tools lack readOnlyHint, destructiveHint, or idempotentHint metadata to help clients understand safety properties.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 58 | 2026-07-28+ | v2 |