Give coding agents a clean way to audit spend.

Xerg exposes one canonical skill, one hosted MCP server, and a small public wire contract so agent tools can run local audits, read recommendations, and query hosted workspace data.

Agent surfaces

SKILL.md

The canonical agent instruction file is served at xerg.ai/skill.md and bundled inside the npm CLI package at skills/xerg/SKILL.md.

Hosted MCP

Pro workspaces can connect remote MCP clients to https://mcp.xerg.ai/mcp with a workspace API key.

CLI

Agents can run local audits through the public npm package without requiring hosted auth.

Wire schemas

TypeScript consumers can depend on the public MIT-licensed @xerg/schemas package for push payload and recommendation contracts.

Client setup

Codex

Use the bundled skill for repeatable audit workflows and configure hosted MCP in ~/.codex/config.toml.

Cursor

Use xerg mcp-setup or project MCP config for hosted recommendations after a pushed audit.

Claude Code and Claude Desktop

Use the standard remote MCP JSON shape with a Bearer token header.

Antigravity and other tools

Use the generic streamable HTTP MCP settings until a verified product-specific import format is documented.