Axf MCP server
{
"mcpServers": {
"smartergpt-axf": {
"command": "npx",
"args": [
"-y",
"@smartergpt/axf@latest"
]
}
}
}{
"mcpServers": {
"smartergpt-axf": {
"command": "npx",
"args": [
"-y",
"@smartergpt/axf@latest"
]
}
}
}{
"mcpServers": {
"smartergpt-axf": {
"command": "npx",
"args": [
"-y",
"@smartergpt/axf@latest"
]
}
}
}{
"servers": {
"smartergpt-axf": {
"command": "npx",
"args": [
"-y",
"@smartergpt/axf@latest"
]
}
}
}{
"mcpServers": {
"smartergpt-axf": {
"command": "npx",
"args": [
"-y",
"@smartergpt/axf@latest"
]
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall npx -y @smartergpt/axfversion 2.1.3
What it does
AXF exposes repository-specific operations (build, test, validation, diagnostics) as stable, inspectable capabilities over MCP. It lets agents discover and run repository workflows without rediscovering how each project is organized, targeted at teams with repeated local ceremony across multiple agent interactions.
- Discover what operations a repository supports, inspect their contracts (arguments, outputs, side effects), and execute them with declared authority boundaries.
- Reduce repeated agent work by encoding repository-specific build, test, and validation routes once as named capabilities that any agent can find and follow.
- Evaluate whether a repository has enough repeated local ceremony to justify fitting it into AXF capabilities, using a structured assessment process.
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.