A firewall for your AI agent's tools — wrap any MCP server with spending limits, action limits, human approval gates, and a full audit log. Zero code, one config change.

YAML mandate (policy) loader and evaluator for the Cord protocol — defines what an agent may do without human approval.

Verifiable spend-control policies for Hedera AI agents — drop-in Hedera Agent Kit hooks for budgets, allowlists, human approval, and an immutable HCS audit trail.

TOTP-authenticated command gateway - intercept dangerous shell commands, require human approval

Human approval gate for AI agent actions. Drop-in protection against prompt injection for any API key.

Local human approval server and page for PermitRail agent tool calls. For demos, single-user dev, and internal tools.

Durable, model-agnostic runtime for AI agents: persistent runs, human approval gates, pre-flight budgets, OpenTelemetry export, deterministic replay, and the open OAR run format. Zero runtime dependencies.

Interven AI firewall — TypeScript SDK. Scan agent tool calls before they execute. Block malicious requests, redact PII/secrets, route risky actions to human approval.

Socializer (socializer.co) — AI-native social media management. API + CLI + MCP control plane, Postgres-backed, multi-profile, with an assistive AI layer behind human approval and an immutable audit log.

Habena — keep your AI agent on a short rein. MCP middleware proxy: policy guardrails, spend caps, and human approval for AI agents.

Pure, tested risk classifiers that decide which AI-agent actions need a human approval on an AgentBee hardware key.

Agent-to-agent relay with human approval gates. E2E-encrypted rooms over a Cloudflare Durable Object relay; every inbound request and outbound response is approved by a human.

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

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

MCP proxy server that intercepts dangerous AI agent tool calls for human approval via Slack or email

n8n community node for Pushary: send push notifications and pause your workflow for human approval from your phone or Slack.

Sandclaw Browser plugin — web research via Confidante with human approval

A verifier for capabilities an author can't author past. Declare a capability manifest, scope its permissions, harden it before approval, replay its execution, and require a hash-bound human approval before it runs.

Make AI videos from a one-line idea — a step-by-step CLI (vclaw) over Veo, Seedance & Runway with a readable JSON artifact at every stage, a human approval gate before any spend, and automated vision QC that catches motion defects (vapour, morphing, vanis

Human-friendly process signals