This MCP server allows running various AI CLI tools in background processes with automatic permission handling. It supports multiple AI models and enables efficient coding task management.
From the registry: MCP server for AI CLI tools (Claude, Codex, Gemini, and Forge) with background process management

$ claude --dangerously-skip-permissionsPlease install the `ai-cli-mcp` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **Node.js 20+** — ESM support required.
Optional prerequisites:
- Claude CLI first-time setup — Run Claude CLI manually once with --dangerously-skip-permissions to accept terms. Run: `claude --dangerously-skip-permissions`
Canonical MCP server config (stdio transport):
- command: `npx`
- args: ["-y","ai-cli-mcp@latest"]
- optional environment variables:
- `CLAUDE_CLI_NAME`: Override Claude CLI binary name or absolute path. (example: `claude`)
- `CODEX_CLI_NAME`: Override Codex CLI binary name or absolute path. (example: `codex`)
- `GEMINI_CLI_NAME`: Override Gemini CLI binary name or absolute path. (example: `gemini`)
- `FORGE_CLI_NAME`: Override Forge CLI binary name or absolute path. (example: `forge`)
- `OPENCODE_CLI_NAME`: Override OpenCode CLI binary name or absolute path. (example: `opencode`)
- `MCP_CLAUDE_DEBUG`: Enable debug logging to stderr. (example: `false`)
Note: Runs Claude/Codex/Gemini/Forge/OpenCode CLIs as background processes with PID tracking.
Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.AI Agents Framework with Self Reflection and MCP support