Clarify the two supported Rig entry points in README

[Copilot]

The README currently centers on a single quick-start path and local development commands, but does not clearly present the two primary ways Rig is intended to be used. This update reframes the doc around those entry points: using Rig as a skill in Copilot SDK workflows, and running Rig programs directly via skills/rig/rig.ts.

• Usage model

  • Replaces the generic quick start with a clear “2 ways to use Rig” structure.
  • Makes the intended choice obvious at the top of the README instead of requiring readers to infer it from deeper docs.

• Copilot SDK skill path

  • Adds a minimal workflow snippet showing how to pin githubnext/rig/skills/rig/SKILL.md.
  • Adds a minimal Rig program example that explicitly configures copilotEngine().

• Direct launcher path

  • Shows both supported launcher modes:
    • piping an inline program to node skills/rig/rig.ts
    • running a program file with stdin
  • Includes the --typecheck path so readers can validate programs without executing them.

• Doc focus

  • Removes the local development section from the README.
  • Narrows the README to entry-point guidance and routes detailed behavior to skills/rig/SKILL.md and skills/rig/references/runtime.md.

import { agent, configureAgent, copilotEngine, p, s } from "rig";

configureAgent(copilotEngine());

const reviewDiff = agent({
  model: "small",
  instructions: p`Review ${p.bash("git diff -- .")} and return only the declared output.`,
  output: s.object({
    summary: s.string,
    risk: s.enum("low", "medium", "high"),
  }),
});

export default reviewDiff;