1import{j as e}from"./vendor-xyflow-Brp8IoVU.js";import{G as n}from"./GeoPageTemplate-B1oVmeum.js";import"./vendor-react-D3Nsd4Dj.js";import"./index-BRjqV66B.js";import"./vendor-supabase-CMEY6uiz.js";import"./arrow-right-DT3OzmND.js";import"./chevron-right-CS41fpfV.js";import"./shield-aux2qftm.js";const o={title:"OpenAI Codex Integration with Runwall â MCP Governance",description:"How to connect OpenAI Codex to Runwall for execution governance via MCP. Enforce policies, risk scoring, and audit logging on Codex agent tool calls.",path:"/integrations/codex",breadcrumbs:[{name:"Home",href:"https://runwall.in/"},{name:"Integrations",href:"https://runwall.in/integrations/codex"},{name:"Codex",href:"https://runwall.in/integrations/codex"}],content:["OpenAI Codex is an autonomous coding agent that can execute multi-step development tasks including reading code, running tests, installing dependencies, and deploying changes. When connected through MCP, Codex's tool calls can be governed by Runwall â ensuring that every file operation, terminal command, and API call is evaluated against policy before execution.","Codex connects to Runwall via the HTTP/SSE transport endpoint. Once configured, Codex's autonomous workflow operates normally for safe actions while high-risk operations are caught by Runwall's governance pipeline â risk-scored, evaluated against policy, and routed for approval when necessary."],codeSnippet:{title:"Codex MCP Connection",language:"text",code:`# Codex HTTP/SSE MCP endpoint 2https://mcp.runwall.in/mcp 3 4# Configure with your Runwall API key in the Codex 5# MCP settings panel or environment configuration. 6# All tool calls will route through Runwall's governance 7# pipeline automatically.`},faqs:[{question:"Does Runwall work with OpenAI Codex?",answer:"Yes. Codex connects to Runwall via the HTTP/SSE MCP endpoint. Once connected, all of Codex's tool calls â file operations, terminal commands, API requests â are governed by Runwall's policy engine, risk scoring, and audit logging."},{question:"Can Runwall govern Codex's autonomous multi-step workflows?",answer:"Yes. Runwall evaluates each tool call independently as it occurs. In a multi-step Codex workflow, every step passes through the governance pipeline. If any step triggers a policy violation or exceeds a risk threshold, that specific action is caught â without disrupting the rest of the workflow."},{question:"How do I set up Codex with Runwall?",answer:"Configure Codex to connect to the Runwall MCP endpoint (https://mcp.runwall.in/mcp) using HTTP/SSE transport. Provide your Runwall API key in the connection configuration. Codex will then route all MCP tool calls through Runwall's governance pipeline."}],relatedLinks:[{label:"Getting Started Guide",to:"/docs/getting-started"},{label:"Claude Code Integration",to:"/integrations/claude-code"},{label:"Autonomous Agents Use Case",to:"/use-cases/autonomous-agents"},{label:"Custom Agents Integration",to:"/integrations/custom-agents"}]};function d(){return e.jsx(n,{data:o})}export{d as default};
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.