"Creating the Best General-Purpose Agent" matching MCP servers:

A Model Context Protocol (MCP) server for producing better Terraform through CLI analysis (tflint, checkov, trivy, kics, infracost), best-practice guidance from terraform-best-practices.com, cloud provider recommendations (Azure, AWS, GCP), and Terraform Registry resource and module guidance.

A local MCP daemon that turns an agentic coding client into a bug bounty operator, with 103 tools for offensive security testing including MITM proxy, traffic analysis, and OOB callbacks.

Screens AI agent tool calls before execution and tool results before the agent reads them, acting as an MCP proxy and hooks adapter with shadow, enforce, and strict policy modes.

Enables AI coding agents to run one-shot commands and copy single files to and from remote servers through The Bastion, which authenticates, authorizes, and records each action. Access is controlled by the bastion rather than allowlists or deny lists.

Enables AI assistants to perform password security auditing using John the Ripper on a remote Kali system via SSH, supporting cracking, hash management, and session control.

the-real-unifi-mcp License: MIT Node MCP SDK [![CI](https://github.com/mguttmann/the-real-unifi-mcp/actions

Local stdio MCP server for Bitwarden (Public API + Secrets Manager). 50+ admin tools for members, groups, collections, policies, licensing, projects, and secrets. For Claude Desktop / Claude Code.

Enables autonomous agents and MCP-compatible clients to enforce policy-approved actions using risk tiers, reversible checkpoints, and token approval bounds.

Provides tools to issue, verify, and export cryptographically signed receipts for AI agent actions, enabling tamper-proof audit trails for compliance with regulations like the EU AI Act.

Registers a set of inert, tempting decoy tools with any MCP-capable client so that any call from a hijacked or prompt-injected agent instantly raises an alert with full attack context and a per-call trace token. Also plants worthless canary tokens in honeypot files that trigger zero-false-positive leak alerts when they surface in agent output, files, or outbound requests.

Verifies agent actions against mandates via a single tool, returning allow, deny, or requires_approval receipts without executing the actions.

Enables an LLM agent to read from and propose changes to a financial ERP, executing writes only after human confirmation delivered through the channel, with role checks, risk tiers, audit logging, and a master kill-switch.

Local static inspection of MCP and AI-agent tool manifests before attachment, reporting findings and missing evidence without executing proposed tools. Results do not establish runtime safety.

Cryptographic accountability for AI agents. Ed25519-signed receipts for every MCP tool call. Constraints, chains, AI judgment, invoicing, and local dashboard included.

Security scanning, prompt injection detection, secret leak detection, and agent permission auditing for AI agent workflows.

Enables AI agents to securely store and retrieve secrets and make authenticated API calls on their behalf, keeping API keys out of conversation context and supporting local file or GCP Secret Manager backends.

MCP server for AI agent guardrails — validates and signs agent decisions (OAA tokens) so third parties can verify them without server access

agent-bom is an open security scanner and self-hosted control plane for AI, MCP, and cloud infrastructure. The default scan profile in MCP server mode exposes 8 MCP tools. Additional profiles provide inventory, findings, compliance, graph, and runtime workflows. See the published Glama release for its packaged version.

Enables coding agents to send files and notes into approved Zas channels and read items back, with content encrypted before leaving the machine and scoped, revocable agent access.

Enables AI agent tool calls to be intercepted and judged before execution, allowing, asking a human, or blocking based on calibrated risk and policy.

Pay-per-call MCP and x402 v2 server on Base USDC providing 19 keyless tools for AI agents: web extraction, social and academic search, B2B lead and domain intelligence, npm/PyPI package and smart-contract security audits, OFAC compliance screening, and DeFi swap quotes

Living off the Land intelligence for AI agents. GTFOBins, LOLBAS, LOOBins, LOLDrivers, LOLRMM, WADComs, LOLC2 and more — unified into a single MCP server.

Cryptographic identity and trust protocol for AI agents. 38 MCP tools across 8 protocol layers: Ed25519 identity, delegation chains, values compliance, signed communication, policy engine, task coordination, cross-layer integration, and agentic commerce. 264 tests passing.

Enables AI agents to send, receive, verify, approve, and delegate signed, proof-carrying requests across organisations using delegation chains, replay protection, quorum approvals, and tamper-evident receipts.

Enterprise-grade code intelligence server for AI assistants that analyzes codebases for due diligence, tech debt, security vulnerabilities, and code quality.

Enables AI agents to control a Linux desktop and browser through governed tools for mouse, keyboard, windows, clipboard, screenshots, navigation, and page interaction, with policy verdicts, human confirmation, secret redaction, and an append-only audit trail.

Enables authenticated clients to call a streamable HTTP MCP endpoint with a whoami tool, providing identity lookup for authenticated sessions.

An MCP server that lets agents act as the person who added them, giving them identity, shared context (data, documents, goals, calendar, skills), and a place to publish apps and pages behind sign-in. It also brokers access to models, data APIs and other tools' MCP servers through one key, with consent prompts, per-person access checks and encrypted credentials as guardrails.

Enables defenders to deploy a decoy MCP tool server that records and fingerprints how LLM agents probe, escalate, and persist, without exposing real systems.

Enables authorized web/API security analysis as a stateful, multi-stage workflow with persistent session state, human-controlled validation, candidate versus confirmed finding tracking, and stop conditions.