Agent Toolbox

io.github.solhammerPublished by GitHub account solhammerv0.1.4Updated Oct 24, 2026

26 deterministic, offline pre-action safety gates for AI agents, each with a signed verdict.

VerifiedSTDIODesktop onlyFinance Developer Tools Security & Monitoring

Server details

Identifier: io.github.solhammer/agent-toolbox
Transport: STDIO
Runtime: Desktop only
Trust: Verified
Published: Sep 16, 2026

Security note

MCP servers receive tool arguments from a conversation and may perform external actions. Review what a server can do before you enable it, and give it only the access it needs.

The registry verified the GitHub account solhammer for this listing. That does not show the account is the product's maker.

SourceWeft's checks cover whether the listing is complete and the endpoint answered; they do not review the server's code or what its tools do.

Add a badge to your README

Show readers this MCP server is listed on SourceWeft. The badge links to this page.

[![Listed on SourceWeft](https://sourceweft.com/badge/mcp/io.github.solhammer%2Fagent-toolbox.svg)](https://sourceweft.com/mcp/io.github.solhammer%2Fagent-toolbox?utm_source=readme&utm_medium=badge)

Overview

AI-generated overview

Agent Toolbox is a local MCP server offering 15 offline pre-action safety and quality checks for AI agents, each returning a signed verdict.

What you can do: Validate code and outputs: check imports, verify output quality, distill context into a smaller form, and count tokens before acting on a result. Run security and privacy scans: detect secrets, prompt injection, vulnerabilities and personally identifiable information. Check finance figures: convert units, look up prices and symbols, assess token risk, slippage, order risk and position exposure. All checks run locally over stdio and return a signed verdict.
When to use it: Use it when you want an assistant to screen its own code, outputs or plans before acting, or to run quick security, privacy and finance sanity checks without calling an external service.
Requirements: A local Node.js runtime to run the npm package over stdio, for example via npx. No API key or environment variables are needed; it relies on free public data sources.

Before you install

The finance tools give risk and price assessments, not advice, and should not be relied on for trading decisions. The security scans are heuristic checks, so a clean verdict does not guarantee safety. Verdicts are signed, but the README does not explain how to verify them.

Installation

In SourceWeft

Open Agent Toolbox in the dashboard and add it to a workspace.
Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

agentoolbox-mcp

MCP server for agent-toolbox.ai — the quality layer for AI agents. Exposes code- and output-quality tools to Claude, Cursor, Warp, Cline, and any MCP-compatible client over stdio.

Tools (15)

Core: validate_imports, verify_output, distill_context Security & privacy: scan_secrets, scan_injection, count_tokens, scan_vulnerabilities, scan_pii Finance: finance_units, finance_price, finance_symbol, finance_token_risk, finance_slippage, finance_order_risk, finance_position_check Runs locally over stdio, in-process — no API key or env vars (free public data sources only).

Install

No clone or build needed:

json

{  "mcpServers": {    "agent-toolbox": { "command": "npx", "args": ["-y", "agentoolbox-mcp"] }  }}

{  "mcpServers": {    "agent-toolbox": { "command": "npx", "args": ["-y", "agentoolbox-mcp"] }  }}

Or build from source:

bash

git clone https://github.com/solhammer/agentoolboxcd agentoolboxpnpm installpnpm --filter "agentoolbox-mcp..." build   # builds the server + its workspace deps

git clone https://github.com/solhammer/agentoolboxcd agentoolboxpnpm installpnpm --filter "agentoolbox-mcp..." build   # builds the server + its workspace deps

This produces packages/mcp/dist/index.js (use "command": "node" with its absolute path).

Configure your MCP client

Claude Desktop: ~/Library/Application Support/Claude/claude_desktop_config.json
Cursor: ~/.cursor/mcp.json
Warp: Settings → Agents → MCP servers → Add

json

{  "mcpServers": {    "agent-toolbox": {      "command": "node",      "args": ["/absolute/path/to/agentoolbox/packages/mcp/dist/index.js"]    }  }}

{  "mcpServers": {    "agent-toolbox": {      "command": "node",      "args": ["/absolute/path/to/agentoolbox/packages/mcp/dist/index.js"]    }  }}

Replace /absolute/path/to/agentoolbox with the path where you cloned the repo.

The full REST API lives at api.agent-toolbox.ai; all 15 tools are exposed over MCP. License: MIT.

Source: packages/mcp/README.md at commit ad24be5

Tools

SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

v0.1.4LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
DescriptionRequired15 pts
Logo10 pts
Documentation15 pts
License10 pts