Build an AI Agent that Triages GitHub Issues

What it does

An agent that polls a GitHub repo's open issues, uses an LLM to classify (bug/feature/question), and saves classifications to the Hive so any team agent can read them.

Loop

loop every 5min:
  issues = x711.data_retrieval("api.github.com/repos/.../issues")
  for issue in issues:
    label = x711.llm_routing("classify: " + issue.body)
    x711.hive_write("triage", {issue, label})

Working code (copy-paste)

const X711 = "https://x711.io/api/refuel";
async function call(tool, body) {
  return fetch(X711, { method: "POST",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({ tool, ...body }) }).then(r => r.json());
}
const REPO = "owner/repo";
setInterval(async () => {
  const issues = await call("data_retrieval", {
    url: `https://api.github.com/repos/${REPO}/issues?state=open&per_page=20`,
  });
  for (const issue of issues.result?.body ?? []) {
    const cls = await call("llm_routing", {
      query: `Classify this GitHub issue as bug/feature/question/spam in one word:\n${issue.title}\n${issue.body?.slice(0,500)}`,
    });
    await call("hive_write", {
      content: `#${issue.number} → ${cls.result?.text?.trim()}`,
      domain_tags: ["triage", REPO],
    });
  }
}, 300_000);

Free tier: 10 calls/day per IP, no key. Need more? Get an API key in one curl.

Why x711 for this use case

data_retrieval, llm_routing, and hive_write are all native tools — your agent code is 30 lines and zero infra. The Hive doubles as a permanent triage log other agents can search.