The Toolbox MCP
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@The Toolbox MCP search for sqlmap and tell me what it does"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
The Toolbox MCP
๐ AI Agent Tool Discovery via Librarian + Tavily
An MCP (Model Context Protocol) server that enables AI agents to discover installed tools and packages using Librarian's system-wide index, enriched with real-time web intelligence from Tavily.
โจ Features
Tool Discovery: Query 9,236+ tools across 14 package managers
Full-Text Search: Millisecond-fast FTS5 queries
Tavily Integration: Real-time tool research and descriptions
Safe: Read-only, bounded, no tool execution
MCP Native: Works with Claude, Bifang, and other MCP clients
Related MCP server: fossick-mcp
๐ ๏ธ Tools Provided
1. toolbox_search
Search for tools by keywords with pagination.
2. toolbox_describe
Get detailed info for a specific tool.
3. toolbox_inventory
Overview of indexed tools, sources, and categories.
4. toolbox_research ๐ (Tavily)
Enrich tool information with real-time web search via Tavily.
๐ Quick Start
Prerequisites
Python 3.11+
uvpackage managerLibrarian index database
Tavily API key (optional, for enrichment)
Setup
# Clone this repo
git clone https://github.com/Protocol-X-01/toolbox-mcp.git
cd toolbox-mcp
# Install dependencies
uv sync --locked
# Set environment variables
export TAVILY_API_KEY='tvly_your_key_here'
# Run the server
uv run --frozen the-toolbox --db /path/to/librarian/index.db๐ Tavily Integration
The Toolbox MCP uses Tavily instead of Bright Data for tool enrichment:
Benefits
โ
Real-time Research: Current documentation and examples
โ
High Quality: LLM-optimized results
โ
Affordable: Pay-per-use, no subscription
โ
Simple: One API key, works everywhere
โ
Transparent: Clear sources and citations
How It Works
When toolbox_research is called:
Search Librarian index for tool
Query Tavily for latest info
Combine indexed + web data
Return enriched result
Example
# Search for a tool
result = toolbox_search("sql injection")
# Enrich with Tavily
enriched = toolbox_research("sqlmap")
# Returns:
{
"name": "sqlmap",
"source": "pacman",
"category": "webapp",
"version": "1.10.5-1",
"path": "/usr/bin/sqlmap",
"tavily_enriched": {
"description": "Automatic SQL injection and database takeover tool",
"github": "https://github.com/sqlmapproject/sqlmap",
"documentation": "https://sqlmap.org",
"last_updated": "2026-10-29"
}
}๐ Stats
Current Index (as of 2026-10-29):
Total Tools: 9,236
Sources: 14 package managers
Categories: 60+ types
Query Time: <8ms average
๐ Integration
With Bifang
export BIFANG_MCP_SERVERS=[{
"name":"the-toolbox",
"command":"/path/to/toolbox-mcp/.venv/bin/the-toolbox",
"args":["--db","/var/lib/librarian/index.db"]
}]With Claude Desktop
{
"mcpServers": {
"the-toolbox": {
"command": "/path/to/toolbox-mcp/.venv/bin/the-toolbox",
"args": ["--db", "/var/lib/librarian/index.db"],
"env": {
"TAVILY_API_KEY": "tvly_your_key_here"
}
}
}
}๐ฏ Why Tavily?
vs Bright Data:
Feature | Tavily | Bright Data |
Setup | 1 API key | Complex proxy |
Quality | LLM-optimized | Raw SERP |
Cost | Pay-per-use | Subscription |
Rate Limits | Generous | Strict |
Maintenance | Easy | Complex |
Agent-Friendly | โ Yes | โ ๏ธ Manual |
Winner: Tavily for AI agents! ๐
๐ง Configuration
Environment Variables
# Required for Tavily enrichment
TAVILY_API_KEY=tvly_xxxxx
# Optional: Custom LangSmith tracing
LANGSMITH_API_KEY=lsv2_pt_xxxxx
LANGSMITH_PROJECT=toolbox-mcp
LANGSMITH_ENDPOINT=https://api.smith.langchain.comIndex Path
The Toolbox reads from Librarian's SQLite index:
# Default path
/var/lib/librarian/index.db
# Or specify custom path
uv run the-toolbox --db /custom/path/index.db๐ API Reference
toolbox_search
toolbox_search(
query: str, # Search terms (1-256 chars)
source: str = None, # Filter by package manager
category: str = None, # Filter by category
limit: int = 15, # Results per page (1-50)
offset: int = 0 # Pagination (0-10000)
) -> dicttoolbox_describe
toolbox_describe(
name: str # Exact tool name (1-256 chars)
) -> dicttoolbox_inventory
toolbox_inventory() -> dicttoolbox_research ๐
toolbox_research(
name: str # Tool name to enrich via Tavily
) -> dict๐ก๏ธ Safety Features
โ
Read-Only: SQLite opened with mode=ro
โ
Bounded Inputs: Max query length, pagination limits
โ
Timeout Protection: 1-second query deadline
โ
No Execution: Discovers tools, doesn't run them
โ
Untrusted Data: Descriptions are evidence, not instructions
๐งช Testing
# Run tests
uv run pytest tests/ -v
# Test Tavily integration
export TAVILY_API_KEY='your_key'
uv run python -c "
from toolbox import toolbox_research
result = toolbox_research('nmap')
print(result)
"๐ฆ Supported Sources
Librarian indexes tools from:
pacman (Arch + BlackArch)
aur (AUR packages)
path ($PATH binaries)
cargo (Rust)
pip/pipx (Python)
npm/yarn/pnpm (JavaScript)
docker (Container images)
desktop (GUI apps)
foundry (Solidity)
conda (Data science)
rpm (RPM packages)
And more...
๐ค Integration with Bifang
Built for the Nebius x NVIDIA Hackathon 2026:
Live Test: Nemotron Nano roundtrip verified
Tokens: 8,885 tracked in live demo
Status: Production ready โ
๐ Use Cases
1. AI Coding Agents
"Find tools for SQL injection testing" โ Returns: sqlmap, jsql-injection, sqlbrute...
2. Security Researchers
"What wireless tools do I have?" โ Returns: aircrack-ng, wifite, reaver...
3. DevOps Engineers
"Find container tools" โ Returns: docker, podman, buildah...
4. Data Scientists
"What ML libraries are installed?" โ Returns: numpy, pandas, scikit-learn...
๐ Links
Repository: https://github.com/Protocol-X-01/toolbox-mcp
Librarian: https://github.com/Protocol-X-01/librarian
Tavily: https://tavily.com
MCP Protocol: https://modelcontextprotocol.io
๐ License
MIT License - See LICENSE
๐ Credits
Librarian - System-wide tool indexing
Tavily - Web intelligence API
MCP - Model Context Protocol
Bifang - AI coding agent integration
Built with โค๏ธ for the Nebius x NVIDIA Hackathon 2026
Available Tools
4 toolstoolbox_describeARead-onlyIdempotent
Return up to 50 indexed records for an exact tool name across package managers.
Inputs: name (required). Returns: object
toolbox_inventoryARead-onlyIdempotent
Show inventory size, source scan timestamps/failures, and categories without scanning or executing tools.
Inputs: no named parameters. Returns: object
toolbox_researchBRead-only
Enrich tool information with real-time web search via Tavily.
Combines Librarian's indexed data with Tavily web search for up-to-date documentation, GitHub links, and usage examples.
Inputs: name (required). Returns: object
toolbox_searchARead-onlyIdempotent
Find installed tools by literal words, with optional source/category filters and bounded pagination.
Inputs: query (required), limit, offset, source, +1 more. Returns: object
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelogโฆ
TDQS
Scored across 4 tools
Each tool has a distinct role: search (fuzzy word lookup), describe (exact-name multi-source lookup), inventory (metadata/status), and research (web enrichment). Search vs describe slightly overlap since both return tool records, but the 'literal words' vs 'exact name' distinction is stated clearly enough to guide selection.
All four tools follow a consistent toolbox_<verb> snake_case pattern (search, describe, inventory, research). Predictable and easy to map to intent.
Four tools is on the lean side but each covers a genuine stage of discovery (scan status, find, detail, enrich). No redundant tools, though the surface feels slightly thin for a 'toolbox'.
For a discovery/lookup server the lifecycle (inventory status -> search -> describe -> research) is covered without dead ends. Missing install/execute/update operations, but those appear outside the stated read-only scope.
Maintenance
Related MCP Connectors
Real-time package and vulnerability data (Python, npm, Cargo, Go, and more) for AI coding agents.
Package intelligence for AI agents across npm, PyPI, crates.io and deps.dev. No API keys.
Package intelligence for AI agents across npm, PyPI, crates.io and deps.dev. No API keys.
GitHub, npm, PyPI, Stack Overflow, arXiv and OpenAlex paper search as agent tools.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to search documentation of packages and services to find implementation details, examples, and specifications.MIT
- AlicenseAqualityDmaintenanceEnables AI agents to prospect across GitHub, PyPI, and npm, searching repositories, packages, and code patterns with read-only tools.780 PyPI5MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to search and retrieve detailed package information from multiple package registries (NPM, crates.io, NuGet, PyPI, Go).MIT
- AlicenseAqualityBmaintenanceProvides software supply-chain intelligence for AI agents, enabling them to query package metadata, versions, downloads, dependencies, and health signals for npm, PyPI, and crates.io packages without API keys.6MIT