MCP Think TankBLOCK
MCP Think Tank is a powerful Model Context Protocol (MCP) server designed to enhance the capabilities of AI assistants like Cursor and Claude. It provides a structured environment for enhanced reasoning, persistent memory, and responsible tool usage.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
[](https://www.npmjs.com/package/mcp-think-tank) [](https://opensource.org/licenses/MIT) [](https://www.typescriptlang.org/) [](https://www.cursor.so/) [](https://www.anthropic.com/) [](https://github.com/modelcontextprotocol) [](https://nodejs.org/) [](https://smithery.ai/server/@flight505/mcp-think-tank)
Overview
MCP Think Tank is a powerful Model Context Protocol (MCP) server designed to enhance the capabilities of AI assistants like Cursor and Claude @Web. It provides a structured environment for enhanced reasoning, persistent memory, and responsible tool usage.
Key capabilities include advanced **Sequential Think
0aca6f8e2f7fOBSERVED · 2026-10-07Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add mcp-think-tank --env EXA_API_KEY=${EXA_API_KEY} -- npx -y [email protected]{
"mcpServers": {
"mcp-think-tank": {
"command": "npx",
"args": [
"-y",
"[email protected]"
],
"env": {
"EXA_API_KEY": "${EXA_API_KEY}"
}
}
}
}Exposed tools (21)
12 read · 6 write · 3 destructive. Blast radius: 3 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
add_observations | write | Add new observations to existing entities in the knowledge graph |
axios | read | HTTP client for improved URL fetching |
complete_task | read | Mark a task as completed. |
create_relations | write | Create multiple new relations between entities in the knowledge graph. Relations should be in active voice |
delete_entities | destructive | Delete multiple entities and their associated relations from the knowledge graph |
delete_observations | destructive | Delete specific observations from entities in the knowledge graph |
delete_relations | destructive | Delete multiple relations from the knowledge graph |
exa_answer | read | Ask a question and get a sourced answer via Exa /answer API. |
exa_search | read | Search the web using Exa API |
list_tasks | read | List tasks with optional filtering by status and priority. |
lru-cache | read | LRU cache for tool call limiting |
memory_query | read | Query the memory store with advanced filters |
next_task | read | Get the next highest priority todo task and mark it as in-progress. |
open_nodes | read | Open specific nodes in the knowledge graph by their names |
plan_tasks | write | Create multiple tasks from a plan. Generates IDs and syncs with knowledge graph. |
read_graph | read | Read the entire knowledge graph |
search_nodes | read | Search for nodes in the knowledge graph based on a query |
show_memory_path | read | Return absolute path of the active knowledge-graph file. |
update_relations | write | Update multiple existing relations in the knowledge graph |
update_tasks | write | Update multiple tasks with new values. |
upsert_entities | write | Create new entities or update existing ones in the knowledge graph using an upsert pattern |
Trust audit
BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | WARN |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (5 observation(s))
- Network
- declared (1 observation(s))
- Shell
- declared (3 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (14)
exec(`npm install ${dep.name}`, (error: Error | null) => {mcp-think-tank-2.0.7.tgz
delete_entities, delete_observations, delete_relations
.eslintignore
.eslintrc.js
const hash = crypto.createHash('sha1');packageJsonPath = resolve(__dirname, '../../package.json');
import { createDirectory } from '../../utils/fs.js';import { IAgent } from '../../agents/IAgent.js';import { IAgent } from '../../agents/IAgent.js';import { BasicAgent } from '../../src/agents/BasicAgent.js';async-mutex, axios, chalk, exa-js, fastmcp, lru-cache, minimist, node-fetch
assets/MCP_Think_Tank_dark.png
assets/MCP_Think_Tank_light.png
Gates applied: no_behavioural_pass.
0aca6f8e2f7ffull audit observations/trust-audit/mcp-server/flight505__mcp-think-tank.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-07 | 0aca6f8e2f7f | BLOCK | D | 69 | first audit |
Questions
What is the MCP Think Tank MCP server?
MCP Think Tank is a powerful Model Context Protocol (MCP) server designed to enhance the capabilities of AI assistants like Cursor and Claude. It provides a structured environment for enhanced reasoning, persistent memory, and responsible tool usage.
What tools does MCP Think Tank expose?
21 in total: 12 read-only, 6 that write, and 3 that can delete or overwrite (delete_entities, delete_observations, delete_relations). Every one is listed on this page with its risk.
Is MCP Think Tank safe to connect to an agent?
No — not without reading the findings first. The audit graded it D (69/100) and found 1 critical or high issue in the source. Each one is listed on this page with the file and line it is on. Separately from the audit: 3 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does MCP Think Tank need?
It reads EXA_API_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does MCP Think Tank run?
It speaks streamable-http, so it runs as a service you connect to over the network. It is published on npm as mcp-think-tank at 2.0.7.
How current is this page?
The grade is for one exact copy of the source (0aca6f8e2f7f), read on 2026-10-07. The repository is watched and re-audited when it changes.