Files
PaperClipAI/scripts/__tests__/grok-public-install-sandbox.test.mjs
T
DottaandPaperclip d67e3507b2 feat(runner): qualify Codex platform packages and cache hosted builds (#16028)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work.
> - Native execution needs a compatible installed daemon and provider
dependency.
> - The Codex-default work found missing platform packages and
clean-install startup failures.
> - Release builds also repeat dependency compilation and run
preparation serially.
> - This prerequisite packages four daemon targets and verifies
installed native startup.
> - It also caches hosted Cargo dependencies and overlaps JavaScript
preparation with compilation.
> - Agent creation defaults and experimental gates stay at the current
master behavior.

## Linked Issues or Issue Description

Refs #15683. Refs #15555. Refs #15422. Refs #15926. Refs #15932.

Extract runtime and installation support from #15683. This prerequisite
has 65 changed files. Creation, onboarding, imports, UI runner selection
and graduation remain in the dependent PR. Feature/access maps and the
Storybook source catalog stay synchronized. Existing agents keep their
execution choices.

## What Changed

- Correct the existing runner test fixtures. Reserve owned credential
quorum listeners through the existing bounded fixture helper. Keep
foreign listeners intact, run actual credential staging once, and
preserve its rejection and unchanged-auth assertions. Reuse the
durable-resume outer-limit correction. Wait for actual warm-completion
and held-authorization callback entry under the existing recovery
deadline and surface early request errors. Preserve every semantic
assertion.
- Retain bounded installation failure diagnostics through the existing
qualification receipt upload. Keep the original error, 10-minute
deadline and all success checks.
- Keep startup ownership until credential preparation and registration
drain. Reject duplicate and post-close startup/resume work. Observe
shutdown during normal and adopted authentication readiness. Preserve
the primary failure and retain private state when cleanup is
unconfirmed.
- Package Linux x64/ARM64 and macOS Apple Silicon/Intel daemons. Verify
target headers, source identities, manifests, and hashes before assembly
and installation.
- Reuse the existing release producer/consumer workflows. Cache Cargo
dependencies by target, OS, compiler, lockfile, and contract inputs.
Rebuild workspace binaries. Permit trusted cache writes only for
canonical exact-source master pushes.
- Overlap JavaScript preparation with native compilation. Join both
commands and failure/cancellation cleanup before artifact verification.
- Preserve shell positional arguments in the shared SSH runner and apply
environment assignments correctly. Reject environment option keys. Six
before/after controls and two real SSH subprocess paths pass with zero
provider calls; this fix lands before both Codex and Claude defaults.
- Admit verified suspended SSH backups during cold controller
continuation, using the same full-scope, integrity and lifecycle checks
as sandbox targets. Two positive SSH cases fail before; all 14
positive/negative matrix cases pass afterward. Active ownership and
corrupt/malformed/nonempty state remain rejected.
- Record measured sandbox platform facts through the existing explicit
environment probe. Bind facts to the company, effective provider
configuration, immutable OCI image and actual captured image binding.
Check the returned cleanup receipt. Reject client-authored proof and
stale or named boot identities. Creation and discovery allocate nothing.
Preserve connection-only behavior.
- Resolve the actual remote Codex target and preserve same-host
development overrides. Preserve runtime artifact verification. Use
actual protocol/model compatibility rather than a numerical CLI version
fence.
- Resolve the installed Codex login dependency on 64-bit Linux and Mac
independently of the daemon libc baseline. Preserve account homes and
safe errors.
- Fix actual clean-install TOML table comparison. Preserve strict
semantic checks, comments, private writes, and malformed-input
rejection. Preserve primary startup failures and bounded cleanup.
- Document GNU libc 2.39+ and stock Ubuntu sandbox prerequisites.
Qualify ARM startup with a scoped distribution profile and unchanged
global restrictions. Preserve Windows and Alpine deferral.
- Preserve selected and saved native Codex reasoning effort in the
existing UI builder, while excluding it from other providers. This small
shared correction prevents import edits higher in the stack from
dropping effort.
- Reuse existing installed-package checks and assertions. Paperclip
tarballs contain no Codex native payloads. Official npm dependencies
supply the consumer's Codex binaries. All Docker and native builds stay
remote.

## Verification

Candidate: `271d434c170e74bdedde65d0b13059a55419c943`. It includes
current master `e168512f7bdbb6af6ea01954c6d7ed80b91e903e` and preserves
prior candidates through ordinary commits.

Current [full
CI](https://github.com/paperclipai/paperclip/actions/runs/38346601686)
is successful on attempt 2. GitHub reports 54 successful checks and four
conditional skips, including required verify and browser aggregates,
workspace typecheck/tests/build, Linux DB controls, both runner shards
and [isolated native
compilation](https://github.com/paperclipai/paperclip/actions/runs/38346601609).
Fresh current-head Greptile is 5/5 with all four threads resolved. The
passing build's clean Grok/Codex report records CI merge source
`7a82a37438e6e448abc7cc5caa60dc2519a97be0`, 18 installed packages,
omitted Codex payloads, official host-only npm dependencies and zero
provider calls. Pi setup and its no-prompt lease verification also pass.
The first attempt's single 60-second Node archive download timeout
remains retained; its DNS/TLS/download cause remains unknown. One
unchanged-source retry was requested through the normal Build-job dialog
with debug logging disabled. GitHub's attempt 2 graph reports the other
CI jobs as well; no second retry or source/deadline relaxation was made.

The [final exact-source package
campaign](https://github.com/paperclipai/paperclip/actions/runs/38347027535)
is successful on literal `271d434c170e74bdedde65d0b13059a55419c943`. It
passes all four native producers, public Linux package assembly and
clean installed consumers on Apple Silicon, Intel Mac and Linux ARM64.
The nine qualification-mode skips are intentional; full CI above covers
their ordinary gates. Native builds and Docker assembly ran remotely.
The predecessor [`dae808` package
campaign](https://github.com/paperclipai/paperclip/actions/runs/38342104102)
remains literal historical evidence. Only tests and the dated record
differ from that source; runtime, package and workflow code is
unchanged. The reused credential fixture correction passes all 42
canonical Node 24 controls with synthetic data and independent
matching-source review, and the current hosted runner shard passes.
Production credential behavior is unchanged.

**Merge hold:** current-source CI, fresh 5/5 review and exact-source
package qualification pass. Actual code-owner approval and normal
protected merge remain pending. No slice has merged or deployed. GitHub
CLI authentication needs refresh; the normal signed-in embedded GitHub
UI provided this status without credential workarounds.

- The master merge preserves both Storybook source-review appendices and
the guarded test-home cleanup. All six existing cleanup controls pass
without skips. Master's isolated-home helper/test/setup are
byte-identical. The two existing whitespace lines in master's unrelated
privacy test remain outside this PR's diff.
- Canonical Node 24 passes all 24 existing public-install helper tests.
Three narrow added cases prove bounded/redacted failure metadata,
symlink exclusions, original exception identity, failed diagnostic
writes and no false failure on success. The Mac installer now uses
`--timing`; at most two 16 KiB tails and 32 structured entries per tail
enter the existing receipt upload. Raw argv, URLs, environment, error
text and stacks are omitted. A failed receipt explicitly remains
FAILED/unqualified/runtime-not-attempted. The 600,000 ms install
deadline and every real package/hash/payload/lock/lifecycle/runtime
oracle are unchanged.
- Canonical Node 24 reproduces the durable-resume outer timeout under
controlled fake latency. Only its outer test limit changes to 15
seconds; internal 500 ms/5 second deadlines and usage/effect assertions
remain. Controlled warm-recovery and before-bootstrap authorization
scheduling delays reproduce their callback polling failures, while the
explicit entry signals pass. Early recovery errors surface unchanged.
Independent review is clear. Historical mock binaries support the
fixture controls; fresh full hosted CI remains decisive. The original
hosted callback failure cause is inconclusive.
- The composed runner/surfaces/server/E2E TypeScript checks pass. Map
checks pass 35 areas/171 capability IDs. Token gates pass. The existing
Storybook catalog passes 223 files/2,210 stories; story IDs and
classifications remain intact. The current master had 25 inherited stale
fingerprints from three changed map recipes. Only those guarded
fingerprints and a source-only review note change; all 92 non-map
story/component/fixture inputs remain unchanged. Existing
classifications are retained. This proves source structure, not visual
or live acceptance.
- The 65-file contribution audit has no code file crossing from below
1,000 lines to above the limit. Earlier comment cleanup preserves
executable emissions, parsed workflows and action pins. Independent
cleanup/catalog and three-file reporting reviews are clear. No new
workflow, suite, grader or local native/Docker build was added.

**Prior `dae808` gate evidence:** fresh Greptile is 5/5 with four
resolved threads, isolated native compilation and the browser aggregate
pass. Its first runner shard passes the previous two corrections but
fails a different held-authorization cell at the implicit one-second
polling deadline. It records 1,672 passes/one failure/three skips. This
candidate applies the verified entry-signal correction without changing
production deadlines or any semantic assertions. Its original hosted
cause remains inconclusive until fresh CI.

**Previous-source evidence, kept literal:** the [exact `16a532` package
campaign](https://github.com/paperclipai/paperclip/actions/runs/38338806076)
passes all four native producers, public Linux package assembly, Apple
Silicon, Intel Mac and Linux ARM64 installed consumers. Its fresh review
is 5/5 with all four threads resolved. Full CI records 52 successful
checks, four conditional skips and the two fixture failures corrected
here. This is not a green full-CI candidate or proof for the new source.

`64c074fb8da1b31f8d6042c4fef48dc4164e759b` passed [full
CI](https://github.com/paperclipai/paperclip/actions/runs/38332726938),
54 checks/four conditional skips, [isolated native
compilation](https://github.com/paperclipai/paperclip/actions/runs/38332726787/job/115085895328),
and fresh Greptile 5/5/all four resolved threads. The hosted transport
suite passed 246 tests and startup trust passed 22, closing that
source's local fake-binary limitation. These are historical after
reconciliation.

The [previous exact-source package
campaign](https://github.com/paperclipai/paperclip/actions/runs/38332921463)
built all four platform daemons and passed public Linux package assembly
plus Apple Silicon and Linux ARM64 installed consumers. **Intel Mac
failed:** its initial scripts-disabled npm install timed out before
lifecycle/runtime assertions. The underlying download/resolution/host
cause is unknown because the old helper deleted the npm logs. This is
the concrete baseline for the diagnostic change. The failed campaign
remains failed; no all-platform signoff or automatic retry is claimed.

The earlier `a200495` [four-target
campaign](https://github.com/paperclipai/paperclip/actions/runs/38283695248)
and [scoped ARM
qualification](https://github.com/paperclipai/paperclip/actions/runs/38291535914)
remain component evidence. Cargo restores were cold misses; no
warm-cache speedup is claimed. Paperclip tarballs contain no Codex
payloads. Official npm dependencies supply Codex on the user's platform.

Current-source full workspace typecheck/tests/build, Linux DB controls,
static/Rust/native tests, all browser shards and exact-source package
consumers must pass. New-agent defaults and final managed onboarding
belong to #15683. Actual private Cloud composition, warm/cold signup,
task/follow-up and rollback compatibility remain later release gates.
Source audits found that ordinary managed onboarding does not run the
saved platform probe and Cloud distributes named snapshots without a
verified final boot-image binding. These dependent release blockers
remain open; prequalified E2E fixtures do not close them. Master adds
task-workspace migrations; the final image/migrator and previous-image
checks must include them. ARM paid task/follow-up and ordinary OS APT
installation are not proven.

## Risks

- Multi-target packages require source and header agreement. Missing or
mismatched artifacts fail explicitly. Existing execution choices remain
unchanged.
- Linux requires GNU libc 2.39+ and compatible sandbox prerequisites.
The scoped hosted ARM preparation does not qualify arbitrary containers
or kernels. Windows, 32-bit ARM, and Alpine remain deferred.
- Only dependency crates are cached. Preview jobs cannot write trusted
caches. Concurrent command cleanup must preserve the primary failure.
- Native-free Paperclip packaging retains official Codex npm
dependencies. No Codex payload bundling, schema migration, automatic
agent migration, or production deployment is included.

## Model Used

OpenAI Codex, GPT-6 family, with reasoning, repository tools, code
execution, and parallel agents. The exact serving model snapshot and
context-window size are not exposed.

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-10-13 06:24:07 -05:00

607 lines
50 KiB
JavaScript

import assert from 'node:assert/strict';
import { createHash } from 'node:crypto';
import { existsSync, mkdirSync, mkdtempSync, readFileSync, realpathSync, rmSync, symlinkSync, writeFileSync } from 'node:fs';
import { spawnSync } from 'node:child_process';
import { tmpdir } from 'node:os';
import { join } from 'node:path';
import test from 'node:test';
import { fileURLToPath } from 'node:url';
import { isWorkspacePackage, parseWorkspaceEntries } from '../build-standalone-public-packages.mjs';
import { GROK_PUBLIC_INSTALL_IMAGE, GROK_PUBLIC_INSTALL_LIFECYCLE, assertInstalledCodexPlatformProvenance, assertInstalledNativeConsumerContinuity, assertNoBundledCodexPayloads, grokConsumerDockerArgs, installedCodexProbeSource, installedDaemonProbeSource, macNodeHeadersDirectory, macPublicInstallLifecycleProfile, macInstalledStartupProbeSource, runMacInitialInstallWithDiagnostics } from '../grok-public-install-sandbox.mjs';
import { verifyPiPackagedDaemon } from '../pi-public-install-probe.mjs';
import { runnerBinaryTarget } from '../../packages/paperclip-runner/src/live/runner-binary.ts';
const paths = { assets: '/private/staging/assets', consumer: '/private/staging/consumer', cache: '/private/staging/cache', uid: 1001, gid: 1001 };
const values = (args, flag) => args.flatMap((value, index) => value === flag ? [args[index + 1]] : []);
test('Mac initial install failures retain bounded npm progress without secret or environment values', () => {
const root = realpathSync(mkdtempSync(join(tmpdir(), 'mac-install-diagnostics-test-')));
try {
const cacheDirectory = join(root, 'cache'), logs = join(cacheDirectory, '_logs'), receiptPath = join(root, 'failure.json');
mkdirSync(logs, { recursive: true });
writeFileSync(join(logs, '0000-debug-0.log'), '0 timing idealTree:init Completed in 987654321ms\n');
const debug = Array.from({ length: 1000 }, (_, index) => `${index} timing idealTree:node_modules/private-secret Completed in ${index}ms`).join('\n');
writeFileSync(join(logs, '0010-debug-0.log'), debug + '\n1001 verbose reify PASSWORD=991122ms\n'
+ '1002 http fetch GET 200 https://user:password@registry.npmjs.org/pkg?token=query-secret 312ms (cache miss)\n'
+ '1003 verbose cli /private/secret/path --token=argv-secret\n1004 silly fetch manifest Authorization=Bearer-header-secret\n');
writeFileSync(join(logs, '0020-timing.json'), JSON.stringify({ timers: { 'idealTree:userRequests': 92967, 'reify:unpack': 76,
'ENV_SECRET': 112233, 'reify:private-secret': 17 }, environment: { PASSWORD: 'environment-secret' } }));
const original = Object.assign(new Error('password=exception-secret'), { code: 'ETIMEDOUT', errno: -60, status: 1, signal: 'SIGTERM',
stdout: 'stdout-secret', stderr: 'stderr-secret', spawnargs: ['--token=argument-secret'] });
const options = { sourceRevision: 'a'.repeat(40), platform: 'darwin', architecture: 'x64', cacheDirectory, receiptPath };
assert.throws(() => runMacInitialInstallWithDiagnostics(() => { throw original; }, options), error => error === original);
const text = readFileSync(receiptPath, 'utf8'), receipt = JSON.parse(text);
assert.equal(receipt.sourceRevision, options.sourceRevision); assert.equal(receipt.platform, 'darwin'); assert.equal(receipt.architecture, 'x64');
assert.equal(receipt.outcome, 'FAILED'); assert.equal(receipt.qualified, false); assert.equal(receipt.runtimeNotAttempted, true);
assert.equal(receipt.phase, 'initial-npm-install'); assert.equal(receipt.providerCalls, 0); assert.equal(receipt.timeoutMilliseconds, 600000);
assert.ok(receipt.elapsedMilliseconds >= 0);
assert.deepEqual(receipt.error, { code: 'ETIMEDOUT', status: 1, signal: 'SIGTERM', errno: -60 });
assert.equal(receipt.npmDiagnosticTail.length, 2);
assert.ok(receipt.npmDiagnosticTail.every(tail => tail.entries.length <= 32));
assert.equal(receipt.npmDiagnosticTail[0].truncated, true);
assert.ok(receipt.npmDiagnosticTail[1].entries.some(entry => entry.phase === 'idealTree:userRequests' && entry.durationMilliseconds === 92967));
assert.ok(receipt.npmDiagnosticTail[0].entries.some(entry => entry.method === 'GET' && entry.status === 200 && entry.durationMilliseconds === 312));
assert.ok(Buffer.byteLength(text) < 10000);
assert.doesNotMatch(text, /secret|PASSWORD|Authorization|https:|private|991122|987654321|112233/i);
} finally { rmSync(root, { recursive: true, force: true }); }
});
test('Mac install diagnostics never follow outside log symlinks or replace the primary failure', () => {
const root = realpathSync(mkdtempSync(join(tmpdir(), 'mac-install-diagnostics-boundary-')));
try {
const cacheDirectory = join(root, 'cache'), logs = join(cacheDirectory, '_logs'), receiptPath = join(root, 'failure.json');
mkdirSync(logs, { recursive: true });
writeFileSync(join(root, 'outside.log'), '1 timing idealTree:userRequests Completed in 987654321ms\n');
symlinkSync(join(root, 'outside.log'), join(logs, '0001-debug-0.log'));
const original = Object.assign(new Error('primary'), { code: 'unsafe-secret', status: null, signal: 'unsafe-secret' });
const options = { sourceRevision: 'b'.repeat(40), platform: 'darwin', architecture: 'arm64', cacheDirectory, receiptPath };
assert.throws(() => runMacInitialInstallWithDiagnostics(() => { throw original; }, options), error => error === original);
const receipt = JSON.parse(readFileSync(receiptPath, 'utf8'));
assert.deepEqual(receipt.npmDiagnosticTail, []); assert.equal(receipt.error.code, null); assert.equal(receipt.error.signal, null);
rmSync(receiptPath); mkdirSync(receiptPath);
assert.throws(() => runMacInitialInstallWithDiagnostics(() => { throw original; }, options), error => error === original);
rmSync(receiptPath, { recursive: true }); rmSync(logs, { recursive: true }); symlinkSync(root, logs);
assert.throws(() => runMacInitialInstallWithDiagnostics(() => { throw original; }, options), error => error === original);
assert.deepEqual(JSON.parse(readFileSync(receiptPath, 'utf8')).npmDiagnosticTail, []);
} finally { rmSync(root, { recursive: true, force: true }); }
});
test('successful Mac installs do not create failed receipts or change the bounded install contract', () => {
const root = realpathSync(mkdtempSync(join(tmpdir(), 'mac-install-diagnostics-success-')));
try {
const receiptPath = join(root, 'receipt.json');
assert.equal(runMacInitialInstallWithDiagnostics(() => 'installed', { sourceRevision: 'c'.repeat(40), platform: 'darwin',
architecture: 'x64', cacheDirectory: join(root, 'cache'), receiptPath }), 'installed');
assert.equal(existsSync(receiptPath), false);
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
assert.match(source, /timeout: 10 \* 60_000/);
assert.match(source, /runMacInitialInstallWithDiagnostics\(\(\) => run\(process.execPath, \[npmCli, 'install', '--ignore-scripts', '--omit=dev', '--timing'/);
assert.ok(source.indexOf('runMacInitialInstallWithDiagnostics(()') < source.indexOf("assert.equal(existsSync(sentinel), false)"));
assert.match(source, /finally \{ rmSync\(root, \{ recursive: true, force: true \}\); \}/);
} finally { rmSync(root, { recursive: true, force: true }); }
});
test('native host startup preserves the consumer lock and both executable byte identities', () => {
const before = { lockSha256: 'a'.repeat(64), runnerSha256: 'b'.repeat(64), codexCommandSha256: 'c'.repeat(64), codexNativeExecutableSha256: 'e'.repeat(64) };
assert.doesNotThrow(() => assertInstalledNativeConsumerContinuity(before, { ...before }));
for (const field of Object.keys(before)) {
assert.throws(() => assertInstalledNativeConsumerContinuity(before, { ...before, [field]: 'd'.repeat(64) }), new RegExp(field));
assert.throws(() => assertInstalledNativeConsumerContinuity({ ...before, [field]: '' }, { ...before, [field]: '' }), new RegExp(field));
}
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
const arm = source.split('async function verifyLinuxArmConsumer')[1].split("if (process.argv[2] === '--linux-arm64-consumer')")[0];
assert.match(arm, /const hostEnv = \{ PATH: .*HOME: home, TMPDIR: hostTmp, LC_ALL: 'C', TZ: 'UTC' \}/);
assert.match(arm, /execFileSync\(process.execPath, \[file\], \{ cwd: consumer, env: hostEnv/);
assert.ok(arm.indexOf('isolated(GROK_PUBLIC_INSTALL_LIFECYCLE)') < arm.indexOf("hostProbe('host-startup-probe.mjs'"));
assert.ok(arm.indexOf('const before =') < arm.indexOf("hostProbe('host-startup-probe.mjs'"));
assert.ok(arm.indexOf("hostProbe('host-startup-probe.mjs'") < arm.indexOf('const after ='));
assert.match(arm, /assertInstalledNativeConsumerContinuity\(before, after\)/);
assert.doesNotMatch(arm, /externallySandboxed|externalSandbox|sudo|sysctl|--privileged|seccomp=unconfined|unshare/);
});
test('Mac offline lifecycle uses only the matching executing Node SDK', () => {
const root = realpathSync(mkdtempSync(join(tmpdir(), 'mac-node-headers-test-')));
try {
const prefix = join(root, 'sdk'), executable = join(prefix, 'bin/node'), include = join(prefix, 'include/node');
mkdirSync(join(prefix, 'bin'), { recursive: true }); mkdirSync(include, { recursive: true });
writeFileSync(executable, 'fixture Node executable');
writeFileSync(join(include, 'node.h'), '#include "node_version.h"\n');
const header = '#define NODE_MAJOR_VERSION 24\n#define NODE_MINOR_VERSION 20\n#define NODE_PATCH_VERSION 0\n';
writeFileSync(join(include, 'node_version.h'), header);
assert.equal(macNodeHeadersDirectory(executable, '24.20.0'), prefix);
assert.throws(() => macNodeHeadersDirectory(executable, '24.19.0'), /must match the executing Node/);
assert.throws(() => macNodeHeadersDirectory(executable, '24.20.0-custom'), /exact release version/);
writeFileSync(join(include, 'node_version.h'), header + '#define NODE_PATCH_VERSION 0\n');
assert.throws(() => macNodeHeadersDirectory(executable, '24.20.0'), /declare one patch version/);
writeFileSync(join(include, 'node_version.h'), header);
rmSync(join(include, 'node.h'));
assert.throws(() => macNodeHeadersDirectory(executable, '24.20.0'), /ENOENT/);
writeFileSync(join(root, 'outside-node.h'), '#include "node_version.h"\n');
symlinkSync(join(root, 'outside-node.h'), join(include, 'node.h'));
assert.throws(() => macNodeHeadersDirectory(executable, '24.20.0'), /inside the executing Node prefix/);
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
assert.ok(source.includes('macNodeHeadersDirectory(process.execPath, process.versions.node)'));
assert.ok(source.includes('consumer, { ...env, npm_config_nodedir: nodeHeaders }'), 'SDK is bound only to the deferred lifecycle environment');
assert.equal(source.split('npm_config_nodedir:').length, 2);
} finally { rmSync(root, { recursive: true, force: true }); }
});
test('lifecycle execution has no network, host credentials, checkout, or elevated privileges', () => {
const args = grokConsumerDockerArgs({ ...paths, command: GROK_PUBLIC_INSTALL_LIFECYCLE });
assert.deepEqual(values(args, '--network'), ['none']);
assert.deepEqual(values(args, '--user'), ['1001:1001']);
assert.ok(args.includes('--read-only'));
assert.deepEqual(values(args, '--cap-drop'), ['ALL']);
assert.deepEqual(values(args, '--security-opt'), ['no-new-privileges']);
assert.deepEqual(values(args, '--mount'), [
'type=bind,src=/private/staging/assets,dst=/packages,readonly',
'type=bind,src=/private/staging/consumer,dst=/consumer',
'type=bind,src=/private/staging/cache,dst=/cache',
]);
assert.deepEqual(values(args, '--env'), ['HOME=/tmp', 'npm_config_cache=/cache', 'npm_config_nodedir=/usr/local', 'npm_config_audit=false', 'npm_config_fund=false', 'npm_config_ignore_scripts=false']);
assert.match(GROK_PUBLIC_INSTALL_IMAGE, /@sha256:[a-f0-9]{64}$/);
});
test('deferred lifecycle execution rebuilds the installed graph without dependency resolution', () => {
assert.deepEqual(GROK_PUBLIC_INSTALL_LIFECYCLE, ['npm', 'rebuild', '--offline', '--ignore-scripts=false', '--dangerously-allow-all-scripts']);
});
test('a root or malformed host identity cannot run lifecycle scripts', () => {
for (const uid of [0, -1, undefined, '1001']) {
assert.throws(() => grokConsumerDockerArgs({ ...paths, uid, command: ['npm', 'ci'] }), /unprivileged/);
}
});
test('only the scripts-disabled dependency download gets network access', () => {
const args = grokConsumerDockerArgs({ ...paths, download: true, command: ['npm', 'install', '--ignore-scripts'] });
assert.deepEqual(values(args, '--network'), ['bridge']);
assert.ok(values(args, '--env').includes('npm_config_ignore_scripts=true'));
});
test('Pi assembly gets bounded scratch capacity while preserving sandbox restrictions', () => {
const args = grokConsumerDockerArgs({ ...paths, download: true, temporarySizeMiB: 2048, command: ['node', '/consumer/node_modules/paperclipai/dist/index.js', 'runtime', 'setup', 'pi'] });
assert.deepEqual(values(args, '--tmpfs'), ['/tmp:rw,nosuid,nodev,noexec,size=2048m,mode=1777']);
assert.deepEqual(values(args, '--memory'), ['3g']);
assert.ok(args.includes('--read-only'));
assert.deepEqual(values(args, '--cap-drop'), ['ALL']);
assert.deepEqual(values(args, '--security-opt'), ['no-new-privileges']);
for (const temporarySizeMiB of [0, 255, 2049, Infinity, '2048']) {
assert.throws(() => grokConsumerDockerArgs({ ...paths, temporarySizeMiB, command: ['node'] }), /bounded/);
}
});
test('only an offline runtime probe can execute its verified scratch snapshot', () => {
const command = ['node', '/packages/pi-public-install-probe.mjs', '/consumer/node_modules/@paperclipai/server'];
const args = grokConsumerDockerArgs({ ...paths, command, temporarySizeMiB: 2048, temporaryExecutable: true });
assert.deepEqual(values(args, '--tmpfs'), ['/tmp:rw,nosuid,nodev,exec,size=2048m,mode=1777']);
assert.deepEqual(values(args, '--network'), ['none']);
assert.deepEqual(values(args, '--user'), ['1001:1001']);
assert.ok(args.includes('--read-only'));
assert.deepEqual(values(args, '--cap-drop'), ['ALL']);
assert.deepEqual(values(args, '--security-opt'), ['no-new-privileges']);
assert.throws(() => grokConsumerDockerArgs({ ...paths, command, download: true, temporaryExecutable: true }), /offline/);
assert.throws(() => grokConsumerDockerArgs({ ...paths, command, temporaryExecutable: 'true' }), /offline/);
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
assert.ok(source.includes("temporarySizeMiB: 2048, temporaryExecutable: true"));
});
test('the separately provisioned executable is exposed read-only to the offline probe', () => {
const prerequisite = '/private/staging/native/grok';
const args = grokConsumerDockerArgs({ ...paths, prerequisite, command: ['node', '/packages/probe.mjs', 'present'] });
assert.deepEqual(values(args, '--network'), ['none']);
assert.equal(values(args, '--mount').at(-1), `type=bind,src=${prerequisite},dst=/opt/paperclip/providers/grok/1.0.13/grok,readonly`);
});
test('verification never elevates PR-controlled provisioning or cleanup on the host', () => {
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
assert.doesNotMatch(source, /\bsudo\b/);
assert.ok(source.includes("const prerequisite = join(root, 'native/grok')"));
});
test('the image and Grok downloads start before staging and join without weakening the npm checks', () => {
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
const linuxSource = source.slice(source.indexOf("assert.equal(process.platform, 'linux'"));
const position = fragment => { const index = linuxSource.indexOf(fragment); assert.notEqual(index, -1, `verifier must contain ${fragment}`); return index; };
// Exactly the pinned consumer image is pulled, for the platform the probes run on.
const imagePull = position("prefetch('image', 'docker', ['pull', '--platform', 'linux/amd64', GROK_PUBLIC_INSTALL_IMAGE])");
// Grok is provisioned into the prefetch directory, never straight into `native/grok`.
const grokProvision = position("prefetch('grok', process.execPath, [join(repo, 'packages/paperclip-runner/scripts/provision-grok.mjs'), stagedGrok])");
assert.ok(source.includes("const stagedGrok = join(prefetched, 'grok')"));
assert.doesNotMatch(source, /provision-grok\.mjs'\), prerequisite\]/);
// Both downloads start before any package is staged and before the first sandbox runs.
const staging = position("visit('@paperclipai/server')");
const firstSandbox = position("isolated(['npm', 'install', '--ignore-scripts', '--omit=dev'");
assert.ok(imagePull < staging && grokProvision < staging);
assert.ok(position("await awaited('image', imageReady)") < firstSandbox);
// The binary moves to the canonical host path only after npm is proven not to
// have provisioned it, after the negative probe, and before the positive probe.
const npmCheck = position("'npm must not provision Grok'");
const negativeProbe = position("isolated(['node', '/packages/probe.mjs', 'missing'])");
const grokJoin = position("await awaited('grok', grokReady)");
const move = position('renameSync(stagedGrok, prerequisite)');
const positiveProbe = position("isolated(['node', '/packages/probe.mjs', 'present'], { prerequisite })");
assert.ok(npmCheck < negativeProbe && negativeProbe < grokJoin && grokJoin < move && move < positiveProbe);
// The children run through the helper covered by prefetch-child.test.mjs, and
// cleanup waits for them before removing the directory they write into.
assert.ok(source.includes("import { startPrefetch, stopPrefetches } from './prefetch-child.mjs'"));
assert.ok(position('await stopPrefetches(prefetches)') < position('rmSync(root, { recursive: true, force: true })'));
assert.equal(linuxSource.match(/rmSync\(root/g).length, 1);
});
test('the bundled-dependency installs run together while the other packages are copied, and pack in order', () => {
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
const position = fragment => { const index = source.indexOf(fragment); assert.notEqual(index, -1, `verifier must contain ${fragment}`); return index; };
// The installs are children of the script's own CLI entry through the helper
// covered by stage-bundled-packages.test.mjs, never in-process calls.
assert.doesNotMatch(source, /prepareBundledPackage/);
assert.ok(source.includes("import { bundledInstallTimings, startBundledInstalls } from './stage-bundled-packages.mjs'"));
// Only the bundled packages' sources are staged before the installs start;
// the installs are registered for cleanup at once, before anything else can
// fail; the other packages are copied while the installs run; the join comes
// before the packs start; and the packs run through the helper covered by
// pack-staged-packages.test.mjs, bundled packages first, each registered for
// cleanup as it starts, with the results put back into dependency order
// before every tarball is checked.
const bundledSources = position('for (const entry of entries.filter(entry => entry.bundled)) stageSource(entry)');
const start = position("startBundledInstalls(entries.filter(entry => entry.bundled), { logDir: prefetched, cwd: repo })");
const registered = position('prefetches.push(...installs.map(({ handle }) => handle))');
const otherSources = position('for (const entry of entries.filter(entry => !entry.bundled))');
const join = position("await awaited('bundled', Promise.all(installs.map(({ handle }) => handle.done)))");
assert.ok(source.includes("import { packConcurrency, packStagedPackages, packTimings } from './pack-staged-packages.mjs'"));
const packs = position("await packStagedPackages([...entries].sort((a, b) => Number(b.bundled) - Number(a.bundled)), { logDir: prefetched, env, concurrency, register: handle => prefetches.push(handle) })");
const reordered = position('packs.sort((a, b) => a.index - b.index)');
const checks = position("for (const tarball of tarballs) assertNoBundledCodexPayloads(run('tar', ['-tzf', tarball])");
const copied = position('for (const tarball of tarballs) {');
assert.ok(bundledSources < start && start < registered && registered < otherSources && otherSources < join && join < packs && packs < reordered && reordered < checks && checks < copied);
const packagingLane = source.slice(position("} else {\nassert.equal(process.platform, 'linux'"));
assert.equal(packagingLane.match(/run\('npm', \['pack'/g).length, 1, 'only the lifecycle sentinel is packed in-process in the packaging lane');
assert.ok(source.includes("destination: join(root, `pack-${index}`)"), 'each pack writes into its own directory');
// The timing line reports the split staging phase, each install, and the
// pack step with what the packs would have cost one after another.
for (const key of ['stage_ui_dist_ms', 'stage_sources_ms', 'stage_bundled_ms', 'stage_pack_ms', 'stage_pack_check_ms']) assert.ok(source.includes(`timing.${key} =`));
assert.ok(source.includes('...bundledInstallTimings(installs)'));
assert.ok(source.includes('Object.assign(timing, packTimings(packs, concurrency))'));
});
test('the verifier stages only workspace packages, so it skips the standalone build release.sh needs', () => {
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
const position = fragment => { const index = source.indexOf(fragment); assert.notEqual(index, -1, `verifier must contain ${fragment}`); return index; };
// The standalone public packages are the ones outside the pnpm workspace
// (the sandbox-provider plugins). release.sh builds them because it publishes
// them; the verifier packs only the closure of the server and the CLI under
// `workspace:` specs, and a `workspace:` spec can only name a workspace
// package, so no tarball here ever contains or depends on a standalone build.
// The one reader of that build is qualification-package retention, which
// packs the Daytona plugin from its `dist`, so the build runs exactly when
// retention is requested: the single reference sits inside that gate, and
// the same flag is what later decides whether retention runs at all.
const standaloneBuild = "run(process.execPath, [join(repo, 'scripts/build-standalone-public-packages.mjs')], repo)";
assert.equal(source.split('build-standalone-public-packages').length - 1, 1, 'the standalone build is referenced once, inside the retention gate');
const gate = position("const retainQualificationPackages = Boolean(process.env.PAPERCLIP_RUNNER_QUALIFICATION_PACKAGES_DIR);\n if (retainQualificationPackages) {\n const standaloneStarted = performance.now();\n " + standaloneBuild);
assert.ok(source.includes('timing.stage_standalone_build_ms ='), 'the gated build is timed under its own key');
assert.ok(source.includes('if (retainQualificationPackages) {\n const retained = retainRunnerQualificationPackages({'), 'retention is gated by the same flag as the build');
assert.doesNotMatch(source, /if \(process\.env\.PAPERCLIP_RUNNER_QUALIFICATION_PACKAGES_DIR\)/, 'no second reading of the retention flag can diverge from the build gate');
const retention = readFileSync(new URL('../retain-runner-qualification-packages.mjs', import.meta.url), 'utf8');
assert.ok(retention.includes("const pluginRelative = 'packages/plugins/sandbox-providers/daytona';") && retention.includes("cpSync(join(repo, pluginRelative, 'dist'), join(plugin, 'dist'), { recursive: true })"), 'retention still packs the Daytona plugin from its dist, which is why the gated build exists');
const repo = new URL('../../', import.meta.url);
const listing = spawnSync(process.execPath, [fileURLToPath(new URL('scripts/release-package-map.mjs', repo)), 'list'], { cwd: fileURLToPath(repo), encoding: 'utf8' });
assert.equal(listing.status, 0, listing.stderr);
const packages = new Map(listing.stdout.trim().split('\n').map(line => line.split('\t')).map(([dir, name]) => [name, { dir, manifest: JSON.parse(readFileSync(new URL(`${dir}/package.json`, repo), 'utf8')) }]));
const needed = new Set();
const visit = name => {
if (needed.has(name)) return;
const entry = packages.get(name); assert.ok(entry, `missing public workspace dependency ${name}`); needed.add(name);
for (const [dep, spec] of Object.entries({ ...entry.manifest.dependencies, ...entry.manifest.optionalDependencies })) if (spec.startsWith('workspace:')) visit(dep);
};
assert.ok(source.includes("visit('@paperclipai/server');\n visit('paperclipai');"));
visit('@paperclipai/server'); visit('paperclipai');
const workspace = parseWorkspaceEntries(readFileSync(new URL('pnpm-workspace.yaml', repo), 'utf8'));
const standalone = [...packages.values()].filter(({ dir }) => !isWorkspacePackage(dir, workspace)).map(({ dir }) => dir);
assert.ok(standalone.length > 0, 'the public package map must still carry standalone packages for this check to mean anything');
assert.ok(standalone.includes('packages/plugins/sandbox-providers/daytona'), 'the Daytona plugin retention packs is a standalone package, so pnpm build alone cannot produce its dist');
for (const name of needed) assert.ok(isWorkspacePackage(packages.get(name).dir, workspace), `${name} is staged by the verifier but is not a workspace package`);
// The server publishes `ui-dist`, so that copy stays, before any package is
// staged, reusing the UI that `pnpm build` produced, and it is timed. The
// gated standalone build keeps release.sh's order: after the closure is
// known and before the UI copy and any staged source.
assert.deepEqual(JSON.parse(readFileSync(new URL('server/package.json', repo), 'utf8')).files, ['dist', 'ui-dist', 'skills']);
assert.ok(source.includes("PAPERCLIP_RELEASE_REUSE_UI_DIST: '1'"));
const uiDist = position("run('bash', [join(repo, 'scripts/prepare-server-ui-dist.sh')], repo)");
assert.ok(position("visit('paperclipai')") < gate && gate < uiDist && uiDist < position('timing.stage_ui_dist_ms =') && uiDist < position('const stageSource = '));
});
test('the installed Codex probe exercises the public export and rejects incomplete or mismatched packages', () => {
const root = realpathSync(mkdtempSync(join(tmpdir(), 'installed-codex-probe-')));
try {
const consumer = join(root, 'consumer'); mkdirSync(consumer);
const target = `${process.platform}-${process.arch}`;
const platformPath = `node_modules/@openai/codex-${target}`;
const platform = join(consumer, platformPath); mkdirSync(platform, { recursive: true });
const nativeVendor = join(platform, 'vendor'); mkdirSync(nativeVendor);
const triple = { 'linux-x64': 'x86_64-unknown-linux-musl', 'linux-arm64': 'aarch64-unknown-linux-musl',
'darwin-x64': 'x86_64-apple-darwin', 'darwin-arm64': 'aarch64-apple-darwin', 'win32-x64': 'x86_64-pc-windows-msvc', 'win32-arm64': 'aarch64-pc-windows-msvc' }[target];
const nativeBytes = Buffer.alloc(64);
if (process.platform === 'darwin') { nativeBytes.writeUInt32LE(0xfeedfacf); nativeBytes.writeUInt32LE(process.arch === 'arm64' ? 0x0100000c : 0x01000007, 4); }
else if (process.platform === 'win32') { nativeBytes.set([0x4d, 0x5a]); nativeBytes.writeUInt32LE(40, 60); nativeBytes.writeUInt32LE(0x00004550, 40); nativeBytes.writeUInt16LE(process.arch === 'arm64' ? 0xaa64 : 0x8664, 44); }
else { nativeBytes.set([0x7f, 0x45, 0x4c, 0x46, 2, 1]); nativeBytes.writeUInt16LE(process.arch === 'arm64' ? 183 : 62, 18); }
const nativeFile = join(nativeVendor, triple, 'bin', process.platform === 'win32' ? 'codex.exe' : 'codex');
mkdirSync(join(nativeVendor, triple, 'bin'), { recursive: true }); writeFileSync(nativeFile, nativeBytes, { mode: 0o755 });
const manifest = { name: '@openai/codex', version: `0.160.0-${target}`, os: [process.platform], cpu: [process.arch] };
writeFileSync(join(platform, 'package.json'), JSON.stringify(manifest));
const entry = { name: manifest.name, version: manifest.version,
resolved: `https://registry.npmjs.org/@openai/codex/-/codex-${manifest.version}.tgz`, integrity: `sha512-${Buffer.alloc(64, 1).toString('base64')}` };
const lock = { lockfileVersion: 3, packages: { [platformPath]: entry } };
writeFileSync(join(consumer, 'package-lock.json'), JSON.stringify(lock));
const server = join(consumer, 'node_modules/@paperclipai/server');
const vendor = join(server, 'dist/vendor/paperclip-runner'), drivers = join(vendor, 'drivers/acpx'); mkdirSync(drivers, { recursive: true });
mkdirSync(join(vendor, 'live'));
writeFileSync(join(vendor, 'live/runner-binary.js'), `export const runnerBinaryTarget = ${runnerBinaryTarget.toString()};`);
writeFileSync(join(server, 'package.json'), JSON.stringify({ name: '@paperclipai/server', type: 'module' }));
const integrity = join(drivers, 'installation-integrity.js');
const security = join(vendor, 'drivers/codex/codex-security-config.js'); mkdirSync(join(vendor, 'drivers/codex'));
const securitySource = `export const codexExecutableReadOnlyRoots = () => [${JSON.stringify(nativeVendor)}];
export const createIsolatedCodexAppServerArgs = (source, roots) => roots.map(root => JSON.stringify(root) + '=\"read\"');`;
writeFileSync(security, securitySource);
const moduleSource = `import assert from 'node:assert/strict'; import { writeFileSync } from 'node:fs';
export function createAcpxPackageJsonResolver(root, manifest) { assert.equal(root, ${JSON.stringify(server)}); assert.equal(manifest, root + '/package.json'); return () => 'selected-authority'; }
export async function verifyQualifiedAcpxInstallation(profile, resolver) { assert.deepEqual(profile, { agent: 'codex', model: 'gpt-5.4' }); assert.equal(resolver(), 'selected-authority');
return { openCommand: async () => ({ close: async () => { writeFileSync(${JSON.stringify(join(server, 'lease-closed'))}, 'ok'); } }) }; }`;
writeFileSync(integrity, moduleSource);
writeFileSync(join(drivers, 'qualified-profiles.js'), "export const resolveQualifiedAcpxProfile = (agent, model) => ({ agent, model });");
const wrapper = join(consumer, 'node_modules/@openai/codex/bin'); mkdirSync(wrapper, { recursive: true });
writeFileSync(join(wrapper, '../package.json'), JSON.stringify({ name: '@openai/codex', version: '0.160.0' }));
const index = join(vendor, 'index.mjs'), command = join(wrapper, 'codex.js'), probe = join(root, 'probe.mjs');
const executable = version => writeFileSync(command, `#!/bin/sh\n[ "$1" = --version ] || exit 65\nprintf '%s\\n' 'codex-cli ${version}'\n`, { mode: 0o755 });
writeFileSync(probe, installedCodexProbeSource(index, consumer, '0.160.0'));
const run = () => spawnSync(process.execPath, [probe], { encoding: 'utf8', timeout: 10_000, env: { PATH: '/usr/bin:/bin', NODE_PATH: '' } });
writeFileSync(index, `export const resolvePinnedCodexCommand = () => ${JSON.stringify(command)};`);
executable('0.160.0');
let result = run(); assert.equal(result.status, 0, result.stderr);
assert.equal(JSON.parse(result.stdout).pinnedCodexCommandVerified, true);
assert.equal(JSON.parse(result.stdout).codexPlatformPackageSource, 'official npm');
assert.equal(JSON.parse(result.stdout).codexConsumerPlatform, target);
assert.equal(JSON.parse(result.stdout).codexCommandSha256, createHash('sha256').update(readFileSync(command)).digest('hex'));
assert.equal(JSON.parse(result.stdout).codexNativeExecutableSha256, createHash('sha256').update(nativeBytes).digest('hex'));
const original = JSON.parse(result.stdout);
nativeBytes[63] = 1; writeFileSync(nativeFile, nativeBytes);
result = run(); assert.equal(result.status, 0, result.stderr);
const changed = JSON.parse(result.stdout);
assert.equal(changed.codexCommandSha256, original.codexCommandSha256, 'native mutation does not change the npm launcher');
assert.notEqual(changed.codexNativeExecutableSha256, original.codexNativeExecutableSha256);
assert.throws(() => assertInstalledNativeConsumerContinuity({ lockSha256: 'a'.repeat(64), runnerSha256: 'b'.repeat(64), ...original },
{ lockSha256: 'a'.repeat(64), runnerSha256: 'b'.repeat(64), ...changed }), /codexNativeExecutableSha256/);
nativeBytes[63] = 0; writeFileSync(nativeFile, nativeBytes);
rmSync(nativeFile);
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /one native executable/);
const foreignNative = join(root, 'foreign-codex'); writeFileSync(foreignNative, nativeBytes, { mode: 0o755 });
symlinkSync(foreignNative, nativeFile);
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /bounded regular file in its selected platform package/);
rmSync(nativeFile); writeFileSync(nativeFile, nativeBytes, { mode: 0o755 });
result = run(); assert.equal(result.status, 0, result.stderr);
assert.equal(JSON.parse(result.stdout).codexQualifiedInstallationVerified, true);
assert.equal(JSON.parse(result.stdout).codexCommandLeaseVerified, true);
assert.equal(JSON.parse(result.stdout).codexNativeSandboxResourcesVerified, true);
assert.equal(readFileSync(join(server, 'lease-closed'), 'utf8'), 'ok');
assert.equal(JSON.parse(result.stdout).providerCalls, 0);
const legacyPath = 'node_modules/legacy-codex-adapter/node_modules/@openai/codex-' + target;
const legacy = join(consumer, legacyPath); mkdirSync(legacy, { recursive: true });
const legacyManifest = { ...manifest, version: '0.156.1-' + target };
writeFileSync(join(legacy, 'package.json'), JSON.stringify(legacyManifest));
lock.packages[legacyPath] = { ...entry, version: legacyManifest.version,
resolved: 'https://registry.npmjs.org/@openai/codex/-/codex-' + legacyManifest.version + '.tgz' };
writeFileSync(join(consumer, 'package-lock.json'), JSON.stringify(lock));
result = run(); assert.equal(result.status, 0, result.stderr);
assert.equal(JSON.parse(result.stdout).codexPlatformPackages.length, 2);
assert.equal(JSON.parse(result.stdout).codexSelectedPlatformPath, platformPath);
assert.equal(JSON.parse(result.stdout).codexSelectedPlatformVersion, '0.160.0-' + target);
const legacyCommand = join(legacy, 'codex');
writeFileSync(legacyCommand, '#!/bin/sh\nexit 0\n', { mode: 0o755 });
writeFileSync(index, `export const resolvePinnedCodexCommand = () => ${JSON.stringify(legacyCommand)};`);
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /Selected Codex platform version must match the qualified pin/);
mkdirSync(join(legacy, 'vendor'));
const legacyNative = join(legacy, 'vendor', triple, 'codex', process.platform === 'win32' ? 'codex.exe' : 'codex');
mkdirSync(join(legacy, 'vendor', triple, 'codex'), { recursive: true }); writeFileSync(legacyNative, nativeBytes, { mode: 0o755 });
writeFileSync(legacyCommand, '#!/bin/sh\nprintf "codex-cli 0.156.1\\n"\n', { mode: 0o755 });
writeFileSync(security, securitySource.replaceAll(JSON.stringify(nativeVendor), JSON.stringify(join(legacy, 'vendor'))));
writeFileSync(probe, installedCodexProbeSource(index, consumer, undefined, { nativeCodex: true }));
result = run(); assert.equal(result.status, 0, result.stderr);
assert.equal(JSON.parse(result.stdout).installedCodexVersion, 'codex-cli 0.156.1');
assert.equal(JSON.parse(result.stdout).codexQualifiedInstallationVerified, undefined);
writeFileSync(probe, installedCodexProbeSource(index, consumer, '0.160.0'));
writeFileSync(security, securitySource);
writeFileSync(index, `import { writeFileSync } from 'node:fs'; export const resolvePinnedCodexCommand = () => {
writeFileSync(${JSON.stringify(join(platform, 'package.json'))}, ${JSON.stringify(JSON.stringify({ ...manifest, altered: true }))});
return ${JSON.stringify(command)}; };`);
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /metadata must match its scanned provenance/);
writeFileSync(join(platform, 'package.json'), JSON.stringify(manifest));
writeFileSync(index, "export const resolvePinnedCodexCommand = () => { throw new Error('Bundled Codex runtime version mismatch'); };");
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /Bundled Codex runtime version mismatch/);
writeFileSync(index, `export const resolvePinnedCodexCommand = () => ${JSON.stringify(command)};`);
writeFileSync(security, securitySource.replace(JSON.stringify(nativeVendor), '"/missing-vendor-root"'));
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /vendor directory must be readable/);
writeFileSync(security, securitySource.replace(JSON.stringify(nativeVendor), `${JSON.stringify(nativeVendor)}, ${JSON.stringify(consumer)}`));
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /must not grant enclosing npm/);
writeFileSync(security, securitySource.replace("'=\"read\"'", "'=\"write\"'"));
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /retain the selected native executable resources/);
writeFileSync(security, securitySource);
const emptySlot = join(server, 'node_modules/@openai/codex-' + target);
mkdirSync(emptySlot, { recursive: true });
result = run(); assert.equal(result.status, 0, result.stderr);
assert.equal(JSON.parse(result.stdout).codexPlatformPackages.length, 2,
'An empty nested npm optional slot must not obscure the qualified hoisted host package');
writeFileSync(join(emptySlot, 'undeclared-native-payload'), 'native fixture; never executed');
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /empty regular directory; found.*undeclared-native-payload/);
rmSync(join(emptySlot, 'undeclared-native-payload'));
writeFileSync(join(emptySlot, 'package.json'), '{ malformed');
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /SyntaxError/);
rmSync(join(emptySlot, 'package.json'));
writeFileSync(integrity, moduleSource.replace('assert.deepEqual(profile,', "throw new Error('qualified executable digest mismatch'); assert.deepEqual(profile,"));
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /qualified executable digest mismatch/);
writeFileSync(integrity, moduleSource.replace("writeFileSync(", "throw new Error('command lease close failed'); writeFileSync("));
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /command lease close failed/);
writeFileSync(integrity, moduleSource);
lock.packages[platformPath].inBundle = true;
writeFileSync(join(consumer, 'package-lock.json'), JSON.stringify(lock));
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /downloaded, not bundled/);
delete lock.packages[platformPath].inBundle;
writeFileSync(join(consumer, 'package-lock.json'), JSON.stringify(lock));
const foreign = join(consumer, 'node_modules/@openai/codex-win32-arm64'); mkdirSync(foreign);
writeFileSync(join(foreign, 'package.json'), JSON.stringify({ name: '@openai/codex', version: '0.160.0-win32-arm64' }));
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /only the host/);
rmSync(foreign, { recursive: true });
executable('9.9.9'); result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /qualified pin/);
rmSync(command); result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /ENOENT/);
const outside = join(root, 'outside'); writeFileSync(outside, '#!/bin/sh\nexit 0\n', { mode: 0o755 });
writeFileSync(index, `export const resolvePinnedCodexCommand = () => ${JSON.stringify(outside)};`);
result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /inside the installed consumer/);
writeFileSync(index, 'export const unrelated = true;'); result = run(); assert.notEqual(result.status, 0); assert.match(result.stderr, /must export the pinned Codex resolver/);
} finally { rmSync(root, { recursive: true, force: true }); }
});
test('actual tar member checks allow Codex JavaScript but reject nested platform packages and native vendor payloads', () => {
const javascript = ['package/package.json', 'package/node_modules/@openai/codex/package.json',
'package/node_modules/@openai/codex/bin/codex.js', 'package/dist/vendor/paperclip-runner/index.js'];
assert.doesNotThrow(() => assertNoBundledCodexPayloads(javascript));
for (const payload of [
'package/node_modules/@openai/codex-linux-x64/package.json',
'package/node_modules/@openai/codex-linux-arm64/package.json',
'package/node_modules/bridge/node_modules/@openai/codex-darwin-arm64/package.json',
'package/node_modules/@openai/codex-win32-x64/package.json',
'package/node_modules/@openai/codex/vendor/x86_64-unknown-linux-musl/codex/codex',
'package/dist/vendor/aarch64-apple-darwin/bin/codex',
'package/dist/vendor/x86_64-pc-windows-msvc/bin/codex.exe',
]) assert.throws(() => assertNoBundledCodexPayloads([...javascript, payload]), /must not bundle/);
assert.throws(() => assertNoBundledCodexPayloads(['package/../outside']), /inside the package/);
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
assert.match(source, /assertNoBundledCodexPayloads\(run\('tar', \['-tzf', tarball\]\)/);
});
test('consumer provenance rejects bundled, foreign, inconsistent and non-official Codex payloads', () => {
for (const target of ['linux-x64', 'linux-arm64']) {
const path = `node_modules/@openai/codex-${target}`;
const manifest = { name: '@openai/codex', version: `0.160.0-${target}`, os: ['linux'], cpu: [target.slice('linux-'.length)] };
const entry = { version: manifest.version, resolved: `https://registry.npmjs.org/@openai/codex/-/codex-${manifest.version}.tgz`,
integrity: `sha512-${Buffer.alloc(64, 1).toString('base64')}` };
const check = (value = entry, installed = [{ path, manifest }]) => assertInstalledCodexPlatformProvenance({ packages: { [path]: value } }, installed, target);
assert.equal(check().codexConsumerHostOnly, true);
const legacyManifest = { ...manifest, version: `0.156.1-${target}` };
assert.equal(check({ ...entry, version: legacyManifest.version,
resolved: `https://registry.npmjs.org/@openai/codex/-/codex-0.156.1-${target}.tgz` }, [{ path, manifest: legacyManifest }]).codexConsumerHostOnly, true);
assert.throws(() => check(entry, [{ path, manifest: { ...manifest, version: `^0.160.0-${target}` } }]), /exact host version/);
for (const change of [
{ inBundle: true }, { link: true }, { version: `0.156.1-${target}` },
{ resolved: `https://example.com/@openai/codex/-/codex-0.160.0-${target}.tgz` },
{ resolved: `https://registry.npmjs.org/@openai/codex/-/codex-0.156.1-${target}.tgz` },
{ resolved: `https://user@registry.npmjs.org/@openai/codex/-/codex-0.160.0-${target}.tgz` },
{ integrity: undefined }, { integrity: 'sha512-AA==' },
]) assert.throws(() => check({ ...entry, ...change }));
assert.throws(() => check(entry, []), /must install its host/);
assert.throws(() => check(entry, [{ path: 'node_modules/@openai/codex-darwin-arm64', manifest }]), /only the host/);
assert.throws(() => check(entry, [{ path, manifest: { ...manifest, name: 'untrusted' } }]), /identity/);
}
});
test('hosted Mac lifecycle denies OS networking and writes outside owned scratch', () => {
const profile = macPublicInstallLifecycleProfile('/Users/runner/work/_temp/owned-check');
assert.match(profile, /\(deny network\*\)/);
assert.match(profile, /\(deny file-write\*/);
assert.match(profile, /subpath "\/Users\/runner\/work\/_temp\/owned-check"/);
assert.throws(() => macPublicInstallLifecycleProfile('relative'), /absolute/);
assert.throws(() => macPublicInstallLifecycleProfile('/tmp/owned\0bad'), /absolute/);
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
assert.match(source, /process.env.GITHUB_ACTIONS, 'true'/);
assert.match(source, /networkProbe, 'allowed'/);
assert.match(source, /networkProbe, 'denied'/);
assert.match(source, /error.code,'EPERM'/);
assert.match(source, /'rebuild', '--offline', '--ignore-scripts=false', '--dangerously-allow-all-scripts'/);
assert.match(source, /readFileSync\(sentinel, 'utf8'\), 'ok'/);
assert.match(source, /readFileSync\(join\(consumer, 'package-lock.json'\), 'utf8'\), lock/);
assert.match(source, /finally \{ rmSync\(root, \{ recursive: true, force: true \}\); \}/);
});
test('installed daemon proof uses the production host resolver and complete source-bound manifest', () => {
const source = installedDaemonProbeSource('/consumer/server/dist/vendor/paperclip-runner', '/consumer', 'a'.repeat(40));
assert.match(source, /defaultCapabilityRunnerdBinary\(\)/);
assert.match(source, /Object.keys\(manifest.platforms\).sort\(\), targets.sort\(\)/);
assert.match(source, /manifest.sourceRevision, "a{40}"/);
assert.match(source, /runnerBinaryTarget\(bytes\), target/);
assert.match(source, /createHash\('sha256'\)/);
assert.match(source, /metadata.binaryContractVersion, expected.contracts.runnerdArtifact/);
assert.match(source, /'linux-arm64'/);
assert.doesNotMatch(source, /PAPERCLIP_RUNNERD_BINARY|runner\/target\/debug|execFileSync\('paperclip-runnerd'/);
});
test('native ARM clean consumers keep the existing container and installed-runtime boundaries', () => {
const args = grokConsumerDockerArgs({ ...paths, platform: 'linux/arm64', command: GROK_PUBLIC_INSTALL_LIFECYCLE });
assert.deepEqual(values(args, '--platform'), ['linux/arm64']);
assert.deepEqual(values(args, '--network'), ['none']);
assert.deepEqual(values(args, '--user'), ['1001:1001']);
assert.ok(args.includes('--read-only'));
assert.deepEqual(values(args, '--cap-drop'), ['ALL']);
assert.deepEqual(values(args, '--security-opt'), ['no-new-privileges']);
assert.ok(values(args, '--mount').some(value => value.endsWith('dst=/packages,readonly')));
assert.throws(() => grokConsumerDockerArgs({ ...paths, platform: 'linux/arm/v7', command: [] }), /supported native Linux platform/);
const source = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
const arm = source.split('async function verifyLinuxArmConsumer(')[1].split("if (process.argv[2] === '--linux-arm64-consumer')")[0];
assert.match(arm, /process.platform \+ '-' \+ process.arch, 'linux-arm64'/);
assert.match(arm, /const platform = 'linux\/arm64'/);
assert.match(arm, /isolated\(GROK_PUBLIC_INSTALL_LIFECYCLE\)/);
assert.match(arm, /installedCodexProbeSource\(installed \+ '\/index.js', '\/consumer', undefined, \{ nativeCodex: true \}\)/);
assert.match(arm, /tests\/runner-e2e\/installed-cli.ts/);
assert.match(arm, /process.versions.node.split\('\.'\)\[0\], '24'/);
assert.match(arm, /macInstalledStartupProbeSource\(consumer, hostServer, daemon.sha256, join\(assets, 'installed-cli.ts'\)\)/);
assert.match(arm, /startupEnvironment: 'native host, same installed graph, isolated empty homes'/);
assert.doesNotMatch(arm, /stage-runner|runner\/target|PAPERCLIP_RUNNER_BINARY|prerequisite:/);
});
test('Mac startup wiring normalizes only validated manifest digests and uses server-scoped tooling', () => {
const source = macInstalledStartupProbeSource('/owned/consumer', '/owned/consumer/server', 'sha256:' + 'a'.repeat(64), '/source/tests/runner-e2e/installed-cli.ts');
assert.match(source, /expectedRunnerSha256:"a{64}"/);
assert.doesNotMatch(source, /expectedRunnerSha256:"sha256:/);
for (const digest of ['a'.repeat(64), 'sha256:bad', 'sha512:' + 'a'.repeat(64), 'sha256:' + 'A'.repeat(64)]) {
assert.throws(() => macInstalledStartupProbeSource('/owned/consumer', '/owned/consumer/server', digest, '/source/helper.ts'), /verified SHA-256/);
}
const verifier = readFileSync(new URL('../verify-grok-npm-install.mjs', import.meta.url), 'utf8');
assert.match(verifier, /\['--import', 'tsx', startupFile\], join\(repo, 'server'\)/);
const closedEnvironment = verifier.match(/const env = \{ PATH: `\$\{tools\}[\s\S]*?\};/)[0];
assert.doesNotMatch(closedEnvironment, /NODE_PATH|NODE_OPTIONS|\.\.\.process.env/);
});
test('Pi admission verifies assembled host daemon identity and preserves the flat source-package control', async () => {
const root = realpathSync(mkdtempSync(join(tmpdir(), 'pi-installed-daemon-'))), server = join(root, 'server');
const output = join(server, 'dist/vendor/paperclip-runner'), target = 'linux-x64', source = 'a'.repeat(40);
const binary = join(output, 'bin', target, 'paperclip-runnerd'), manifestPath = join(output, 'bin/release-manifest.json');
const bytes = Buffer.alloc(64); bytes.set([0x7f, 0x45, 0x4c, 0x46, 2, 1]); bytes.writeUInt16LE(62, 18);
const sha256 = value => `sha256:${createHash('sha256').update(value).digest('hex')}`;
const manifest = { schema: 'paperclip.runner.release-binaries.v1', sourceRevision: source,
platforms: { [target]: { path: `${target}/paperclip-runnerd`, sha256: sha256(bytes) } } };
const check = () => verifyPiPackagedDaemon(server, binary, target, runnerBinaryTarget);
try {
mkdirSync(join(output, 'bin', target), { recursive: true });
writeFileSync(binary, bytes); writeFileSync(join(server, 'dist/build-info.json'), JSON.stringify({ commit: source }));
writeFileSync(manifestPath, JSON.stringify(manifest));
await check();
for (const changed of [null, { ...manifest, schema: 'untrusted' }, { ...manifest, sourceRevision: 'b'.repeat(40) },
{ ...manifest, platforms: { [target]: { ...manifest.platforms[target], path: '../outside' } } },
{ ...manifest, platforms: { [target]: { ...manifest.platforms[target], sha256: 'sha256:' + '0'.repeat(64) } } }]) {
writeFileSync(manifestPath, JSON.stringify(changed)); await assert.rejects(check());
}
const wrongHeader = Buffer.from(bytes); wrongHeader.writeUInt16LE(183, 18);
writeFileSync(binary, wrongHeader);
writeFileSync(manifestPath, JSON.stringify({ ...manifest, platforms: { [target]: { ...manifest.platforms[target], sha256: sha256(wrongHeader) } } }));
await assert.rejects(check(), /linux-arm64/);
writeFileSync(binary, bytes); writeFileSync(manifestPath, JSON.stringify(manifest));
const outside = join(root, 'outside'); writeFileSync(outside, bytes);
rmSync(binary); symlinkSync(outside, binary); await assert.rejects(check());
rmSync(binary); writeFileSync(binary, bytes); rmSync(manifestPath);
const flat = join(output, 'bin/paperclip-runnerd'); writeFileSync(flat, bytes);
await verifyPiPackagedDaemon(server, flat, target, runnerBinaryTarget);
await assert.rejects(check(), /strictly equal/);
writeFileSync(manifestPath, 'null');
await assert.rejects(verifyPiPackagedDaemon(server, flat, target, runnerBinaryTarget), /manifest must be an object/);
} finally { rmSync(root, { recursive: true, force: true }); }
});