CLI for PostWall email security gateway - AI agent email with human approval

MCP server letting an AI agent submit actions for multi-party human approval via Byliner. Cannot approve, reject, or delegate — by design.

Sandclaw Browser plugin — web research via Confidante with human approval

GitOps-driven chaos engineering pipeline with human approval gate

Multi-AI workflow framework — structured plan → cross-review → human approval → execution pipeline for AI coding agents

Human approval guardrails for agent-controlled repositories, deployments, and production-facing actions.

OKed plugin for OpenClaw. Gates skill and tool calls behind a human approval push to your phone.

Thin TypeScript client for RelayKit — let any agent pause for human approval, input, review, or takeover

Generate an MCP server from your Prisma schema — typed per-model tools instead of execute_sql, destructive writes gated behind human approval, every action on an audit log. Plus the operator surface: approvals, audit verify, status, studio.

Agent-safe Obsidian vault access for Pi: auto-detect, retrieve, validate, plan, write, edit, manage, and explicitly destroy Markdown with human approval.

Human approval workflows, deterministic review tickets, escalation, delegation, and audit receipts for Ajnas agent infrastructure.

Belay — a local-first guardrail for AI coding agents (a Lunara product). Intercepts an agent's database/infra actions, computes the blast radius, blocks destructive ops until keychain-gated human approval, and writes a cryptographically signed audit log

The governed ReAct loop: deny-by-default tools, L0-L4 guardrails, human approval pause/resume, full audit chain.

n8n community node for ClearedBy — gate an action behind policy + human approval before your workflow proceeds.

DSH approval/completion chime: distinct sounds, an input-row badge, and system notifications (only while the page is hidden) for human-approval, question, and task-completion events.

Run a marketing org on a fleet of small scoped AI agents, with one human approval gate in the middle.

Route LLM-classified intents through a deterministic registry with a human approval gate

MCP server that proposes Signal messages and groups but sends nothing without explicit human approval via a local web GUI.

Interface Design Gate for Pi plus contract-first skill: pause and require human approval of contract/API shape before implementation lands.

Give AI agents a wallet with a spending limit. MCP server + typed client for Veyra: non-custodial USDC payments on Base with per-payment caps, daily caps, human approval bands and a recipient allowlist enforced server-side.