M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
T

T3MP3ST MCP Server

Community
A multi-agent red-teaming platform for authorized security testing.
How to install ↓ Compare GitHub source repository ↗
Category
Dev Tools #243 of 438
Stars
★ 6.6k Very popular
Runtime
Node.js · Docker
Credentials
Optional API key
License
AGPL-3.0
Last commit
Tools
1
51FMRS · D

Based on the supplied README, T3MP3ST exposes one MCP tool, security_recon, as part of an offensive-security framework intended for authorized testing. No server manifest, client configuration, transport, or compatibility evidence was supplied, so those fields remain unpopulated.

Strongest · Documentation 15/20 Weakest · Security and permissions 5/20

Reliability
8/20
Security and permissions
5/20
Maintenance
12/20
Documentation
15/20
Setup experience
11/20
Why each score
Reliability 8/20
The README describes an MCP server, security_recon, HTTP APIs, tool counts, and scope controls, but the server manifest is unavailable. The supplied material contains no verifiable CI, committed tests, or evidence of MCP initialization/tool-list handshakes. The happy path is plausible, but actual behavior, error handling, and dependency control cannot be confirmed; under static calibration the score stays below 12.
Security and permissions 5/20
This is explicitly an offensive-security platform whose architecture includes Exploiter, Infiltrator, Exfiltrator, Ghost, credential storage, and credential harvesting, with real network tooling and external-target capability. The README claims authorized-use restrictions, mission-based egress scoping, gated opt-in dangerous tools, and human-sent disclosure drafts. However, those controls are not verifiable from a manifest or source files, and credential handling, data boundaries, confirmation behavior, and defaults remain insufficiently evidenced, warranting a low score.
Maintenance 12/20
The repository is not archived, the supplied metadata reports only 2 open issues, and the README identifies AGPL-3.0, an updater, protected paths, and release-check documentation. This indicates maintenance intent, but there is no evidence here of commit cadence, formal releases, dependency-update practice, accountable maintainers, or a security-response channel, so the score is well below full marks.
Documentation 15/20
The README supplies installation paths, War Room and CLI usage, Docker instructions, offline-model configuration, environment variables, the MCP entry point, API routes, capability status, benchmark caveats, and linked operational documentation. It clearly distinguishes stable, experimental, and planned features. Deductions apply because the manifest and tool parameter schemas are absent, while error handling, costs, permission/data flows, and troubleshooting evidence are incomplete; many capability claims remain README-only.
Setup experience 11/20
The material gives npm, Docker, and direct MCP-server paths and covers local agents, API keys, Ollama, and OpenAI-compatible servers, including localhost binding. It lacks a complete mainstream-client MCP configuration example, runtime-version requirements, dependency prerequisites, and cross-platform verification. Setup also depends on a local agent, credentials, or additional security tools, making the path manual and environment-sensitive.

Static review · not runListed 2026-08-14

Read the FMRS scoring method →

Fit and risk

What it can accessRuns commands or codeUses the network

Best for

  • Security teams operating under written authorization and defined rules of engagement
  • Red-team researchers and security educators
  • MCP users able to manage a local Node.js environment

Not for

  • Targets without explicit written permission
  • Users seeking unrestricted remote scanning or execution
  • Users expecting fully validated end-to-end multi-agent exploitation

Required permissions

  • Access to network and security-testing resources within the authorized scope
  • A local Node.js environment, installed dependencies, and any configured agent or model access
  • The supplied material does not specify MCP-client permissions or authentication

Risks and side effects

  • This is an offensive-security tool and must be used only against owned or explicitly authorized systems
  • Incorrect scoping may cause unauthorized access, data loss, or legal exposure
  • The README says coordinated multi-agent exploitation remains unproven and unreliable
  • External security tools and optional arsenal components may affect targets

Setup

Before you start

Runtime：Node.js · Docker

OPENROUTER_API_KEY optionalsecret Optional OpenRouter API key used as a model provider; only needed if not using a connected local coding agent.
VENICE_API_KEY optionalsecret Optional Venice API key as an alternative provider.
ANTHROPIC_API_KEY optionalsecret Optional Anthropic API key as an alternative provider.
OPENAI_API_KEY optionalsecret Optional OpenAI API key as an alternative provider.
XAI_API_KEY optionalsecret Optional xAI API key for Grok Build (grok-build-0.1), a coding model with native tool-calling.
Other optional settings (5)
TEMPEST_LOCAL_BASE_URL optional Base URL of a local OpenAI-compatible model server (Ollama default http://localhost:11434/api, LM Studio http://localhost:1234/v1) for fully offline operation.
TEMPEST_LOCAL_MODEL optional Local model name (e.g. llama3), used together with TEMPEST_LOCAL_BASE_URL.
T3MP3ST_LOCAL_AGENT_TIMEOUT_MS optional Optional per-call timeout in milliseconds for local agent CLI invocations.
T3MP3ST_TASK_TIMEOUT_MS optional Optional timeout in milliseconds for mission tasks.
T3MP3ST_GENERAL_TIMEOUT_MS optional Optional timeout in milliseconds for planning requests.

From the repository root, run npm install. When working from a Git clone, run npm run build if needed. Then run node dist/mcp-server.js. The supplied material does not include a client-specific configuration example.

Check that it works

Start the MCP server with node dist/mcp-server.js; the security_recon tool should appear in your MCP client's tool list. Calling it against an authorized target and receiving tool-backed recon output confirms the connection.

Troubleshooting

  1. Confirm npm install was run from the repository root
  2. If using a Git clone, confirm npm run build was run when required
  3. Confirm the files needed by node dist/mcp-server.js exist
  4. Verify that the target is explicitly authorized and review scope controls
  5. Consult the repository's MCP Guide and related documentation for additional details

Things to try

Once connected, you can ask your AI assistant things like:

  • Run security_recon against example.com and enumerate its open ports and service fingerprints.
  • Use security_recon to do DNS and HTTP fingerprinting on my authorized target host.
  • Recon 10.0.0.5 (a server I own) and summarize the network findings.
  • Do a black-box external-attacker recon pass on this web app with security_recon and report the results.

Tools 1

security_recon read-only
Provides security reconnaissance capabilities; the supplied material does not specify its detailed inputs, outputs, or permission model.

Use cases

Reconnaissance of owned or explicitly authorized targets
Red-team research, security education, and CTF support
Exposing security reconnaissance to MCP-aware agents

Overview

T3MP3ST is an autonomous red-teaming platform and multi-agent offensive-security framework for authorized targets. Its README states that node dist/mcp-server.js exposes the security_recon tool to MCP-aware agents.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 63.3k · Tools 2 Compare with this →
Compare with other MCP servers

Source revision 29824d5625ed Data synced 2026-10-19 Read the FMRS scoring method