- Conduid
- Marketplace
- #least-privilege
MCP servers tagged least-privilege
38 live MCP servers tagged "least-privilege", ranked by trust score. Tags come from package metadata and repository topics, so this list covers servers across every category that work with least-privilege.
Cap
Codebase Awareness Protocol (CAP) is designed for AI coding agents to reliably understand the codebase
samvas-codes/dawshund_mcp
An MCP server to enumerate AWS IAM data, analyze effective permissions, and visualize access relationships across users, roles, and resources. Built for cloud security engineers w…
Attenu Guard
In-process permission enforcement for AI-agent delegation chains: a sub-agent never holds more than its parent's authority, with an offline-verifiable audit ledger. Python + TypeS…
Strongroom
Own your agent secrets — your agent holds a scoped, single-use lease, never the raw API key. strongroom injects the real secret only at egress, so a leaked prompt or poisoned tool…
Attenu Guard Ts
Attenu Guard for TypeScript — per-agent permissions on every tool call and sub-agent handoff (a sub-agent never holds more than its parent) with a hash-chained audit log verified…
Pod
pod — least-privilege compiler for AI agents. Compile a least-privilege policy from what your agent actually did, then prove what happened.
Mia
Managed Identity for Agents. Scoped, revocable, live-checked identity and enforcement for AI agents acting on behalf of a human.
Scopelens
Your agent needs to read code. Why can it write? Evidence-aware GitHub permission reviews and static MCP configuration checks.
Web Agent Gateway
Local least-authority gateway connecting Web AI clients to user-approved developer resources.
Cap
What can this MCP server actually do — and has it changed since you trusted it? Inspect its declared capabilities, seal them in a lock file, and get told when a later version can…
Two Key
Authorization and execution control for AI agents — independent policy and judge approval before consequential tool actions execute. MCP gateway, capability tokens, auditability,…
Umbra
occlusion for agent CLIs, APIs, and MCP servers. Declare what an agent may run, and everything you did not name stays unreachable.
Opshaven
A least-privilege MCP server for safe, structured VPS inspection and controlled operations over SSH.
Grantguard
Prove a Postgres role cannot perform the writes it is not allowed to perform. Catalog-based least-privilege audit + CI gate + MCP server.
Planner Guard
A decision-layer guardrail for MCP agents. Gates tool calls by intent and logs the planner's reasoning. Complements MCP firewalls, doesn't replace them.
Ghl MCP Scoped
Least-privilege policy wrapper for any GoHighLevel MCP server. Declare which tools an AI agent may call and under what argument conditions; everything else is refused and logged.…
Privilege Profiler
Task-scoped least-privilege security proxy and Wazuh detection bridge for MCP-based AI agents
Policy Forge
Least-privilege MCP tool policy generator, validator, risk scorer, diff reporter, and CI gate.
Local MCP Toolbox
Secure, read-only MCP tooling for local environment inspection with policy controls, redaction, and auditing.
Google MCP Multi Account
Multi-account Google Workspace (Gmail, Drive) for LLM agents — 100% local, default-deny policy, human-in-the-loop access. A macOS MCP server.
Blueclaw
A POSIX-isolated agent host: a Go daemon that runs an AI agent harness on behalf of the person who asked, executes every tool call as that person's own unprivileged Linux user, ho…
Capability Ledger
Infer least-privilege MCP tool policies from agent traces
Twin Governed MCP
Least-privilege MCP boundary and governed Runner receipts for the Twin REST API
Agentcapdiff
AI agent security policy-as-code for detecting capability, permission, and least-privilege changes across MCP, OpenAI, Claude, LangChain, LangGraph, and CrewAI tools.
Agentperms
Least-privilege permissions for AI agents and MCP tools. Record what your agent does, auto-generate a safe policy, and block dangerous tool calls before they happen.
Steward
Open-source identity & access governance for AI agent fleets — effective-access analysis, segregation-of-duties, least-privilege enforcement, and a tamper-evident audit ledger.
Permission Linter
CLI and GitHub Action-ready linter for MCP permissions, secrets, destructive tools, and least-privilege policy.
Barzel Central Gateway MCP Server
Enterprise Model Context Protocol governance gateway and MCP control plane for AI agent tool routing, identity-aware policy, risk scoring, approvals, routebooks, workflow simulati…
Hardened Terminal MCP
Security-hardened MCP server for terminal access: deny-by-default policy engine, shell=False sandboxed execution, cwd jail, secret redaction, fail-closed audit log.
Parimit
Proposal-only payment-intent governance for AI agents: deterministic policy, demo human review, hash-linked audit evidence, REST, MCP, and mock outcomes—without a live payment rai…
Openclaw Assistant
Hardened, self-hosted AI assistant on the OpenClaw gateway: Claude over Telegram with least-privilege mail, drive, calendar, GitHub, and web browsing, behind a default-DENY egress…
SharpPyxis.SqlServer.SchemaMcp
Local, read-only MCP server that lets an AI agent read the structure of a SQL Server database — objects, code, dependencies, columns — and never its data.
Agent Gateway A2a
Reference architecture: governed least-privilege access for multi-agent systems — agent gateway, A2A boundary, policy-as-code, fail-closed (HLD/LLD + threat model)
Protonmail AI
Security-first Proton Mail automation for AI agents — direct QR session-fork auth, local-first Rust CLI + MCP, least-privilege tools, and auditable sessions.
Eunox
A capability firewall for MCP servers: a deny-by-default proxy that enforces which tool calls an AI agent may make (and with which arguments), backed by a signed audit log. One st…
Leasebroker
Local-first broker issuing time-bounded, least-privilege capability leases to AI agents and MCP servers — signed, scoped, expiring; deny-by-default with audit and human veto.
Claude MCP Allow
Generate least-privilege Claude Code permission rules for MCP tools from their live annotations: readOnlyHint tools go to allow, everything else to ask, one mcp__server__tool rule…
Catalog
Task-scoped authority for AI agents