Integrations
Wherever your agents work, your rules go with them
Every integration says plainly what it can prevent. Only tested ones are offered today; the rest are on the way, shown, not hidden.
Available now
Coding agents and GitHub
Claude Code and Cursor
PlannedCoding agents on developers' machines, stopped before a destructive command runs.
PreventsGitHub
PlannedEvery change an agent proposes is checked against your rules before it can be merged.
Stops the consequenceOpenAI Codex
PlannedChecked before it acts on a machine; checked at the pull request for cloud tasks.
PreventsSoftware your team builds
Agent frameworks
PlannedLangGraph, Vercel AI, OpenAI Agents, Claude Agent SDK, CrewAI, Google ADK: one line for your developers.
Prevents when the agent asks firstMCP tools
PlannedOne set of rules for every tool an agent reaches through the Model Context Protocol.
PreventsServers and workstations
PlannedAgents running as services or scripts on Linux and Windows.
Prevents when the agent asks firstClouds
Amazon Web Services
PlannedBedrock AgentCore, checked before AWS's own policy runs.
PreventsCloudflare
PlannedOne-click gateway; agents built on Cloudflare check first.
PreventsGoogle Cloud
PlannedAgents built with Google's kit check before every action.
PreventsMicrosoft Azure
PlannedEvery agent's Microsoft identity on its record; checks before each action.
PreventsBusiness platforms and automation
ServiceNow
PlannedEvery agent that reaches ServiceNow held to one change-control rule.
Prevents (via MCP)Microsoft Copilot Studio
PlannedA check step for the actions that matter; records for the rest.
Prevents when the agent asks firstSalesforce Agentforce
PlannedGate the actions that matter; record everything the agent does.
Records; prevents when wrappedZapier, n8n, Make
PlannedOne check step before the actions your automations take.
Prevents when the agent asks first