Composition

aex composes an agent from explicit per-session inputs. The SDK materializes local bytes before the session lands and the platform mounts them into the managed runtime before the first agent turn.

Need	Primitive
Instructional skill bundles	`Skill.fromDir`, `Skill.fromUrl`, `Skill.fromFiles`
Agent instructions	`AgentsMd.fromPath`, `AgentsMd.fromContent`
Reference files and folders	`File.fromPath`, `File.fromBytes`
Remote tools	`McpServer.remote`, `McpServer.fromId`
Non-secret runtime settings	`environment.variables`, `environment.packages`, `environment.networking`
Runtime secrets for your code	`Secret.value`, `Secret.ref`, `environment.secrets`

import { AgentsMd, File, McpServer, Models, Secret, Skill } from "@aexhq/sdk";

await aex.run({
  model: Models.CLAUDE_HAIKU_4_5,
  message: "Use the attached docs and tools to produce a report.",
  agentsMd: [await AgentsMd.fromContent("Follow the repo conventions.", { name: "repo-rules" })],
  files: [await File.fromPath("./input")],
  skills: [await Skill.fromDir("./skills/report-writer", { name: "report-writer" })],
  mcpServers: [McpServer.remote({ name: "github", url: "https://example.com/mcp" })],
  environment: {
    secrets: { INTERNAL_API_TOKEN: Secret.value(process.env.INTERNAL_API_TOKEN!) },
    networking: { mode: "limited", allowedHosts: ["api.example.com"] }
  },
  apiKeys: { anthropic: process.env.ANTHROPIC_API_KEY! }
});

Secrets stay out of reusable configs. Provider keys go in the top-level apiKeys map; reusable or per-run values for your own code go in environment.secrets. Your code then makes normal HTTP calls with the standard client for that service.