stack/packages/harness/tests/bundle.test.mjs

import { test } from "node:test";
import assert from "node:assert/strict";
import { createHash } from "node:crypto";
import { mkdirSync, readFileSync, statSync, writeFileSync } from "node:fs";
import { join } from "node:path";
import { BundleError, buildBundle, sessionModel, SOURCES } from "../src/bundle.mjs";
import { resolvedFor, scratch } from "./helpers.mjs";

const sha = (p) => createHash("sha256").update(readFileSync(p)).digest("hex");
const mode = (p) => statSync(p).mode & 0o777;

function build(t, resolved, extra = {}) {
  const dir = scratch(t);
  const workspace = join(dir, "ws");
  mkdirSync(workspace);
  const bundle = buildBundle({
    dir: join(dir, "bundle"),
    resolved,
    contract: "# Contract\n\nDo the work.\n",
    tracker: true,
    credentials: [{ service: "gitea", date: "2099-01-01", state: "valid" }],
    session: { run: "run-1", launchedBy: "pm", workspace, toolSocket: join(dir, "tools.sock"), turnMarker: join(dir, "turn.done") },
    piVersion: "0.85.1",
    claudeVersion: "2.1.0 (Claude Code)",
    ...extra,
  });
  return { dir, workspace, ...bundle };
}

test("sessionModel: agent vars win, then the system's execution settings", (t) => {
  assert.deepEqual(sessionModel(resolvedFor(t, "pm")), { harness: "pi", provider: "zai", model: "claude-opus-5-5", thinking: null });
  assert.deepEqual(sessionModel(resolvedFor(t, "coder")), { harness: "pi", provider: "zai", model: "glm-5.3-flash", thinking: null });
  // The test system's adapter is "mock", which isn't a session harness.
  assert.throws(() => sessionModel(resolvedFor(t, "reviewer")), (e) => e instanceof BundleError && /harness "mock"/.test(e.message));
});

test("a pi bundle: prompt, policy, tools and manifest, 0600 in a 0700 directory", (t) => {
  const b = build(t, resolvedFor(t, "pm"));
  assert.equal(mode(join(b.dir, "bundle")), 0o700);
  for (const f of Object.values(b.files)) assert.equal(mode(f), 0o600, f);
  assert.deepEqual(Object.keys(b.files).sort(), ["manifest", "policy", "prompt", "tools"]);
  const prompt = readFileSync(b.files.prompt, "utf8");
  assert.match(prompt, /^# Contract\n\nDo the work\.\n\n## This session\n/);
  assert.match(prompt, /Role instance: pm \(definition pm\)\. Run: run-1\./);
  assert.match(prompt, /Bus tools: send_message, launch, /);
  // Row 90 (#1584): PM fault 10 (#228 c91).
  assert.match(prompt, /^- Only the human stops a run, with `mosaic stop`\. If you think a run should stop, raise a role\.revoke decision with blocking false/m);
  assert.deepEqual(JSON.parse(readFileSync(b.files.policy, "utf8")), {
    harness: "pi",
    workspace: b.workspace,
    tools: [...b.manifest.builtins],
    typed: b.tools.map((x) => x.name),
    network: "api-only",
    needs: ["gitea", "vikunja"],
    credentials: [{ service: "gitea", date: "2099-01-01", state: "valid" }],
  });
  const m = b.manifest;
  assert.equal(m.harness, "pi");
  assert.equal(m.model, "claude-opus-5-5");
  assert.equal(m.piVersion, "0.85.1");
  assert.equal(m.claudeVersion, null);
  assert.equal(m.gateTimeout, null);
  assert.deepEqual(m.skills, []);
  assert.equal(m.reliesOn.length, 5);
  assert.ok(m.typed.includes("launch"));
  assert.deepEqual(Object.keys(m.sources).sort(), ["gate", "piExtension", "tools"]);
  assert.equal(m.sources.piExtension.sha256, sha(SOURCES.piExtension));
  for (const [name, digest] of Object.entries(m.files)) assert.equal(digest, sha(join(b.dir, "bundle", name)), name);
  assert.deepEqual(JSON.parse(readFileSync(b.files.manifest, "utf8")), JSON.parse(JSON.stringify(m)));
});

test("a claude-code bundle adds the wrapped gate hook and the MCP config", (t) => {
  const r = resolvedFor(t, "coder", (d) => (d.roles.coder.vars = { harness: "claude-code", model: "claude-sonnet-5-5" }));
  const b = build(t, r);
  assert.deepEqual(Object.keys(b.files).sort(), ["manifest", "mcp", "policy", "prompt", "settings", "tools"]);
  const settings = JSON.parse(readFileSync(b.files.settings, "utf8"));
  const hook = settings.hooks.PreToolUse[0];
  assert.equal(hook.matcher, "*");
  assert.equal(hook.hooks[0].timeout, 20);
  assert.equal(hook.hooks[0].command, `timeout -k 2 10 '${process.execPath}' '${SOURCES.claudeGate}' '${b.files.policy}' || exit 2`);
  const mcp = JSON.parse(readFileSync(b.files.mcp, "utf8"));
  assert.deepEqual(mcp, { mcpServers: { mosaic: { type: "stdio", command: process.execPath, args: [SOURCES.mcpServer, b.files.tools, join(b.dir, "tools.sock")] } } });
  const m = b.manifest;
  assert.equal(m.piVersion, null);
  assert.equal(m.claudeVersion, "2.1.0 (Claude Code)");
  assert.deepEqual(m.gateTimeout, { n: 10, k: 2, hook: 20 });
  assert.ok(m.builtins.includes("Bash"));
  assert.ok(m.typed.every((n) => n.startsWith("mcp__mosaic__")));
  assert.deepEqual(Object.keys(m.sources).sort(), ["claudeGate", "gate", "mcpServer", "tools"]);
});

// Row 89 (#1583): the prompt names the built-ins as the session's harness names them.
test("the prompt's built-in tools line is the harness's own names, as the manifest lists them", (t) => {
  const line = (b) => readFileSync(b.files.prompt, "utf8").split("\n").find((l) => l.startsWith("- Built-in tools: "));
  const pi = build(t, resolvedFor(t, "pm"));
  assert.equal(line(pi), "- Built-in tools: read, write, edit, grep, find, ls.");
  const claude = build(t, resolvedFor(t, "pm", (d) => (d.roles.pm.vars = { ...d.roles.pm.vars, harness: "claude-code", model: "claude-opus-5-5" })));
  assert.equal(line(claude), "- Built-in tools: Read, Write, Edit, Grep, Glob.");
  for (const b of [pi, claude]) assert.equal(line(b), `- Built-in tools: ${b.manifest.builtins.join(", ")}.`);
  for (const harness of ["pi", "claude-code"]) {
    const none = build(t, resolvedFor(t, "pm", (d) => (d.roles.pm.vars = { ...d.roles.pm.vars, harness, model: "claude-opus-5-5", "limits.tools": [] })));
    assert.equal(line(none), "- Built-in tools: none.", harness);
  }
});

// Row 91 (#1585): the prompt names each project clone, and why a kept one wasn't updated.
test("the prompt names the project clones in the workspace, and none when there are none", (t) => {
  const line = (b) => readFileSync(b.files.prompt, "utf8").split("\n").filter((l) => l.startsWith("- Project clones: "));
  const plain = build(t, resolvedFor(t, "pm"));
  assert.deepEqual(line(plain), []);
  const dir = scratch(t);
  const session = (checkouts) => ({ run: "run-1", launchedBy: "pm", workspace: join(dir, "ws"), toolSocket: join(dir, "tools.sock"), turnMarker: join(dir, "turn.done"), checkouts });
  const two = build(t, resolvedFor(t, "pm"), { session: session([
    { project: "stack", path: join(dir, "ws", "stack"), outcome: "updated", branch: "main" },
    { project: "site", path: join(dir, "ws", "site"), outcome: "kept", reason: "dirty" },
  ]) });
  assert.deepEqual(line(two), [`- Project clones: stack at ${join(dir, "ws", "stack")}, on branch main; site at ${join(dir, "ws", "site")}, not updated at this launch (dirty). Each is a git clone of the project. The host fetches from its origin at each launch; the origin is outside this sandbox.`]);
  const prompt = readFileSync(two.files.prompt, "utf8");
  assert.ok(prompt.indexOf("- Workspace: ") < prompt.indexOf("- Project clones: "));
  assert.deepEqual(line(build(t, resolvedFor(t, "pm"), { session: session([]) })), []);
});

// Row 93 (#1587): the review loop's prompt lines, for the roles with its tools.
test("the prompt explains the review loop to a role with push_working, and reviews to one with fetch_review", (t) => {
  const lines = (b) => readFileSync(b.files.prompt, "utf8").split("\n").filter((l) => /^- (Review loop|Reviews): /.test(l));
  const dir = scratch(t);
  const session = { run: "run-1", launchedBy: "pm", workspace: join(dir, "ws"), toolSocket: join(dir, "tools.sock"), turnMarker: join(dir, "turn.done"), checkouts: [{ project: "stack", path: join(dir, "ws", "stack"), outcome: "cloned", branch: "main" }] };
  const coder = lines(build(t, resolvedFor(t, "coder"), { session }));
  assert.equal(coder.length, 1);
  assert.match(coder[0], /^- Review loop: commit on a branch of your clone, push_working it, then request_review/);
  assert.match(coder[0], /closes only on an approve verdict for its latest push/);
  const reviewer = lines(build(t, resolvedFor(t, "reviewer", (d) => (d.roles.reviewer.vars = { harness: "pi" })), { session }));
  assert.equal(reviewer.length, 1);
  assert.match(reviewer[0], /^- Reviews: a REVIEW-REQUEST names a request id\. fetch_review it/);
  assert.deepEqual(lines(build(t, resolvedFor(t, "pm"), { session })), []);
  // Without a clone neither git tool exists, so neither line is there.
  assert.deepEqual(lines(build(t, resolvedFor(t, "coder"))), []);
});

test("a bundle is written once: an existing file refuses", (t) => {
  const dir = scratch(t);
  mkdirSync(join(dir, "bundle"));
  writeFileSync(join(dir, "bundle", "prompt.md"), "old");
  assert.throws(() => buildBundle({
    dir: join(dir, "bundle"),
    resolved: resolvedFor(t, "pm"),
    contract: "c",
    session: { run: "r", launchedBy: "pm", workspace: dir, toolSocket: join(dir, "s"), turnMarker: join(dir, "m") },
  }), /EEXIST/);
  assert.equal(readFileSync(join(dir, "bundle", "prompt.md"), "utf8"), "old");
});

test("a path with a single quote can't go into the hook command", (t) => {
  const r = resolvedFor(t, "coder", (d) => (d.roles.coder.vars = { harness: "claude-code", model: "claude-sonnet-5-5" }));
  const dir = scratch(t);
  assert.throws(() => buildBundle({
    dir: join(dir, "it's"),
    resolved: r,
    contract: "c",
    session: { run: "r", launchedBy: "pm", workspace: dir, toolSocket: join(dir, "s"), turnMarker: join(dir, "m") },
  }), (e) => e instanceof BundleError && /single quote/.test(e.message));
});