Automatic Codex plan review in Claude Code; on-demand in local MCP hosts; human approval stays yours

Execution Authority for AI agents: agents propose, Decionis independently decides ALLOW / ESCALATE / BLOCK, Presence verifies human approval on escalation, and a SafeExecutor runs only on a single-use intent-bound grant. Every decision leaves a Decision D

Human approval for agent work, with code diff and Markdown review.

PagerDuty routing adapter for human approval delivery.

MCP server that runs agent tool calls (shell, files, HTTP, …) through Runestone Gatekeeper — policy enforcement, human approval, and audit for any MCP client

Moona — 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.

Reference human-approval orchestration for REQUEST_CONFIRMATION flows — pluggable channels (webhook, Slack, Teams, email) + a replay-safe resume via adapter-core confirm().

WeMessage: an open-source macOS gateway that lets AI agents draft iMessages behind a human approval gate. Placeholder pending v1 release.

TypeSafe-powered human approval gate for risky Pi bash commands

A policy layer in front of your AI agent's payments. Set spending limits, require human approval over a threshold, and get a tamper-evident log of every decision. Zero third-party dependencies.

Deny-by-default policy layer for WebMCP tools: registration takeover, human approval, audit timeline.

Dainamite Billing — lightweight recurring-billing engine for Open Mercato (Bill Runs, one-time / recurring / usage items, draft invoices in core/sales with human approval).

A CLI that audits and generates WebMCP tool registrations for a website, verified by an isolated agent driving a real browser, with human approval required before any tool goes live

MCP server that lets AI agents make purchases with human approval. Local-first payments, encrypted credentials, budget controls.

MCP server that lets Claude order food on DoorDash — search, compare real fee-included totals, build carts, and place orders behind a human approval dialog.

AgentMemo MCP server — persistent memory and human approval for AI agents

MCP server for Agent Pocket — remote human approval bridge for AI agents

Serve an ontology registry as an MCP server: typed read tools, destructive writes staged for human approval, and an audit log of every call.

OKed SDK - human approval layer for AI agents

Safe Shopify theme changes for AI coding agents: staging theme, validation, preview, human approval, rollback.