Plugins
Buying-intent scoring for AI agents. Score any text — Reddit, HN, RSS, email, Slack, Discord, support tickets — through a three-judge panel that separates real buyers from sellers, competitors and noise, and hands your agent what it needs to draft the reply. MCP-native. The judgement runs on us, the sending runs on you.
World AgentKit-backed HITL approvals for OpenClaw protected tools
Sapience AI's 6-in-1 middleware suite for OpenClaw. These middlewares sit between your agent and and the systems it can affect — intercepting every turn and every tool call, enforcing Human-in-the-Loop gates, redacting PII, blocking prompt injection, routing to the right model, and capping token spend. All locally. Zero telemetry. Nothing leaves your machine.
Personal agent authority for OpenClaw: decide what your agent may do on its own, what it must ask about first, and what it may never do.
WebMCP (Web Model Context Protocol) — browser-native API for agent-ready websites, enabling structured tool registration, declarative forms, human-in-the-loop commerce, and MCP-B polyfill integration.
Step into your OpenClaw agent's browser from your phone for a CAPTCHA, login or 2FA, then hand it back. One-use private links, no app or VPN.
Tools for AI agents to post jobs and hire humans on Clawhand. Pay workers in USDC on Base.
CTRLRun execution safety for OpenClaw: every tool call checked before it runs, and recorded after.
Phone approvals for OpenClaw. Pushary asks you before your AI agent runs a gated tool call, and blocks the call unless you approve it.
Pauses risky OpenClaw actions, like rm -rf, force-push, destructive SQL and prod deploys, until you approve them in Telegram or on the OKed web dashboard. Everything else just runs. Every decision gets a signed receipt.
