feat: default new Apple Silicon Claude agents to Paperclip Runner (#15993)

## Thinking Path

> - Paperclip lets users choose an agent harness and account.
> - The server resolves the runner for new agents.
> - Claude still needs the same ordinary native-default journey as
Codex.
> - Existing agents must retain their runner, model and account.
> - This PR defaults new Apple Silicon Claude agents to Paperclip Runner
with Opus 5.5.
> - Users can connect and work without a runner override.

## Linked Issues or Issue Description

Refs #15683. This PR depends on #15994 and precedes #15997. Related
scope: #15422, #13978 and #15224.

## What Changed

- Resolve new Apple Silicon Claude agents to `paperclip_runner`,
`provider: acpx`, `acpxAgent: claude`. Keep explicit Legacy, models,
provider routing and saved execution.
- Use `claude-opus-5-5` for ordinary Claude defaults and retain explicit
overrides.
- Reuse existing API-key/subscription connections, credential storage
and checks. Generalize the shared Advanced runner control with Claude
branding.
- Carry configured effort into the shared v9 contract. Preserve
compatible permissions, instructions, skills, timeouts and session
behavior. Require Legacy for unsupported settings.
- Qualify a new immutable managed Claude target at explicit Connect
through the existing permission-gated probe. Use fresh scoped discovery
for test and hire. Discard stale results and require an explicit Connect
after settings change.
- Retain merged Codex onboarding recovery, auth preparation,
import/effort corrections and Grok Bot gates.
- Check the Claude first-task wizard's saved default directly. Update
configuration/runtime documentation and exact Opus 5.5 accounting.

- Align hire fixtures with the merged host-environment restriction.
Native agent-authenticated hires cannot inject host env settings; both
creation endpoints verify that no extra agent is saved.

## Verification

- Updated onto master `9179d8dcc3e908c217fde4f63c9d08ccf8132eeb` through
#15994. Native input v9 carries effort while preserving v7 home modes
and v8 Muse. Codex defaults remain unchanged.
- Focused local verification passed 285 server tests, 323 UI tests, 124
setup/registry tests, and two v9 permission-pin transport cases. Direct
server/UI TypeScript checks passed. An initial 30-second hook timeout
was resolved by serial verification with a longer hook budget: all 122
credential/permission cases and 232 snapshot/environment cases pass.
Direct server TypeScript checking also passes.
- Ordinary web creation with the experiment disabled saved native
Claude, Opus 5.5 and High effort without runner/model overrides. The
actual product task and follow-up both succeeded on integrated source
`b2c9fe43f9cd3bc5dd9d22552da0148845ba71a4` with the cloud-built daemon
from `04e0202ca81387b17cbe7fd094a8338648356fe9`. The final master
reconciliation also changes shared runtime source. Historical proof
retains its source identity; a final-source cloud qualification is
running before refreshing the preview daemon.
- Both actual browser downloads were checked: v1 has three sections/six
checkboxes (451 bytes); v2 preserves the original body and adds two
Rollback checks (612 bytes). The original attachment stayed
byte-identical. The follow-up records session resume and the same
provider session. Total estimated API usage: $0.518432800, with no
pending/unpriced usage in the successful fixture.
- Fixed the live-discovered setup incompatibility: native API-key
validation no longer depends on the global Legacy Claude CLI. The
existing provider endpoint verifies the key without a paid prompt;
subscription/custom routing probes remain intact.
- Fixed security review's native-hire credential redirection:
agent-supplied URLs, proxy/launch settings and other host environment
controls are rejected before inheritance. Native credential references
remain supported and partial edits use the saved harness. Added
create/hire/update regression coverage.
- Current-head cloud CI is green. All three stack PRs have fresh
Greptile 5/5 reviews and zero unresolved review threads. Required
typecheck, test and build gates ran in cloud CI; Docker work stayed
remote. Earlier failed checks remain in their original runs.
- Docker and native release builds ran in cloud infrastructure. An
inadvertent local Rust prerequisite was stopped; no local Docker build
was run.

- The final master merge preserves task-loading, permission,
skill-retention and OpenRouter changes. Pi uses a fresh combined profile
27, preserving both historical profile-24 declarations; Copilot remains
profile 22. Focused protocol checks passed 155 cases (one skipped), the
updated Pi attestation check passed 18 (one skipped), and 91
evaluation/catalog checks passed.

- Latest-application manual continuation passed in 37.942 seconds on
`695e5c005a85bddc8711d3fcddcaf6b7e1fb64d0`, with High effort, native
Opus 5.5, Anthropic API attribution, the same resumed Claude session, an
independently checked 635-byte file and persisted Done status. It uses
the earlier verified Mac daemon; the preceding-source package
qualification passed at `dca62c7aeb4b45f0dd91a8e18d386a16a2658ef8`. That
package run predates the small TypeScript startup fix and is not
relabeled as qualification for it.
- Upgrade limitation: changed compiled sidecar identity rejected old
durable launch state and recovery hit session-open bounds before
eventually resuming. That attempt was operator-cancelled and retained;
the successful manual continuation does not qualify seamless upgrades.

- Local latency profile, Opus 5.5/High unchanged: three alternating
real-ACP loopback samples per variant reduce median provider-ready time
from 5.081s to 4.034s. The matched real API task startup improves from
5.533s to 4.348s, but total reply time changes from 17.933s to 20.221s
because agent work varies. This establishes a startup saving of about
one second, not a consistent total-reply speedup. A real candidate
follow-up passed in 20.146s and resumed the same saved session. The
loopback fixture makes no external model requests; failed initial probes
remain retained.
- The startup change passes 107 focused runtime/effort/lease tests, 53
profile checks (one skipped), TypeScript checking and the
TypeScript-only runtime build. Native hire fixture checks pass 70 tests;
their first local run retained a socket-hangup failure and stale
OpenCode expectation, both accounted for before the passing rerun.
Broader required checks run in cloud CI.

## Risks

- Missing packages, authentication rejection, unsupported settings and
unavailable models must fail explicitly. Existing agents retain saved
execution; explicit runner changes use existing revision/session
invalidation.
- Subscription follow-up remains unverified; remaining live acceptance
uses API keys per human steering. Historical accounting holds and the
separate $50 ceiling remain intact.
- The subscription acceptance limit remains explicit for review. No
Claude database migration is introduced.

## Model Used

OpenAI Codex, GPT-6 family, with reasoning, repository tools, code
execution and browser testing. The exact serving model ID and
context-window size are not exposed in this session.

## 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
- [ ] Focused local tests and final cloud verification 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>

tests/runner-e2e/README.md

@@ -1086,13 +1086,12 @@ model. Configured and provider-observed model identities are reported separately
 There are thirteen cases on `legacy-codex`, `legacy-claude`, `runner-codex`, and
 `runner-acpx-claude`, local only (52 cells). `runner-codex` selects OpenAI in the
 production wizard, omits the runner override, and verifies the saved native
-configuration without changing runners afterward. `legacy-codex` selects
-Legacy runner in Advanced before authentication and creation. Both are labeled
-`production-wizard`. Claude's creation default is unchanged: `legacy-claude`
-retains its wizard configuration, while `runner-acpx-claude` explicitly changes
-runtime through the public API before its first task and is labeled
-`post-onboarding-runtime-switch`. Persona, managed instructions, skills, seeded
-question, and task invocation are preserved. The native path removes the legacy
+configuration without changing runners afterward. `runner-acpx-claude` does the
+same for Claude on its qualified default targets and inspects the saved ACPX
+Claude profile directly. Both Legacy profiles select Legacy runner in Advanced
+before authentication and creation. All four paths are labeled
+`production-wizard`; there is no post-onboarding runner switch. Persona,
+managed instructions, skills, seeded question, and task invocation are preserved. The native path removes the legacy
 Paperclip operational skill because Runner supplies that contract through its
 protocol; other assigned skills, including `/first-task`, are retained.
 Each isolated onboarding company has a $5 task budget. Setup probes have a
@@ -1454,9 +1453,10 @@ prompts, onboarding defaults, or provider permissions are changed.
 
 For native onboarding qualification, select `first-task` with profiles
 `runner-codex,runner-acpx-claude` (26 cells). Codex verifies the production
-creation default with the experimental flag disabled. Claude still uses the
-existing public-API runtime switch before provider work; that case does not
-qualify a native Claude creation default.
+creation default with the experimental flag disabled. Claude also inspects its
+wizard's saved native default with the experimental flag disabled; select that
+profile on a qualified default target. Retained results from the former API-switch
+definition do not establish that the current creation-default journey passed.
 
 Current proof and remaining decisions are recorded in
 [the 21 September qualification report](QUALIFICATION-2026-09-21.md). In particular,
@@ -2222,6 +2222,22 @@ always preserves recovery state regardless of this optional diagnostic flag.
 
 The `file-edit-validate` fixture independently downloads the exact active artifact work product from the tested run. It requires the matching run-attributed attachment, filename, MIME type, recorded byte count and SHA-256, then compares the downloaded content to the expected bytes. A workspace file alone cannot satisfy this gate. Explicit failed-case diagnostic retention follows the final result after integrity, isolation and evidence checks, including incomplete publication.
 
+
+## Claude default rollout acceptance
+
+The Claude first-task onboarding case asserts the wizard's saved native profile
+directly. It must not patch a Legacy agent into native execution after creation.
+Ordinary native defaults are qualified by execution target: the first rollout
+enables Apple Silicon only. API-key acceptance requires a useful downloadable
+file, a follow-up revision, independent downloads of both versions, persisted
+configuration and session continuity after reload/restart. Subscription setup
+uses the existing connection tools and receives separate task/follow-up proof.
+
+Keep exact source, runtime executable digests and reports for each candidate.
+Effort or runtime changes require affected acceptance reruns. No automatic paid
+retry campaign is permitted under the separate $50 Claude qualification ceiling.
+Docker and expensive Rust builds run in cloud CI, never on the local Mac.
+
 ## Projectless task workspaces (explicit-only)
 
 `task-workspaces` contains exactly three cells: legacy OpenCode locally and native

tests/runner-e2e/catalog.ts

@@ -1466,7 +1466,7 @@ export const runnerSuites: readonly RunnerSuiteFixture[] = [
     groups: ["onboarding"],
     profiles: runnerProfiles.filter(profile => ["legacy-codex", "legacy-claude", "runner-codex", "runner-acpx-claude"].includes(profile.id)),
     environments: [localEnvironment], tasks: firstTaskTasks, expectedMatrixSize: 52,
-    definitionMetadata: { version: 5, runGrading: "evidenced-nonexecution-and-refusal", credentialPersistenceCheck: false, questionChoiceMinimum: 2, nativeSetup: "codex-production-default-claude-explicit-runtime-switch", productionInstructions: true, qualityGrading: "informational" },
+    definitionMetadata: { version: 7, runGrading: "evidenced-nonexecution-and-refusal", credentialPersistenceCheck: false, questionChoiceMinimum: 2, nativeSetup: "qualified-platform-wizard-saved-default", claudeDefaultTarget: "darwin-arm64", legacySetup: "explicit-advanced-runner-choice", productionInstructions: true, qualityGrading: "informational" },
   },
   {
     id: "agent-chat", label: "Persistent Agent Chat",

tests/runner-e2e/first-task-flow.ts

@@ -8,7 +8,6 @@ import { captureFirstTaskAttachments } from "./first-task-attachments.js";
 import { waitForFirstTaskReply } from "./first-task-replies.js";
 import { observeCompletionUpdate } from "./completion-update-flow.js";
 import {
-  firstTaskNativeRuntimePatch,
   provisionFirstTaskFixtures,
   isManagedFirstTask,
   startFirstTaskWithRuntimeBound,
@@ -329,7 +328,7 @@ export async function setupFirstTaskFixtures(input: {
     exact: true,
   });
   const savedKey = page.getByRole("combobox", { name: "Saved API key" });
-  if (execution.profile.id === "legacy-codex") {
+  if (["legacy-codex", "legacy-claude"].includes(execution.profile.id)) {
     await page.getByText("Advanced", { exact: true }).click();
     await page.getByRole("button", { name: "Runner", exact: true }).click();
     await page.getByRole("option", { name: "Legacy runner", exact: true }).click();
@@ -359,7 +358,7 @@ export async function setupFirstTaskFixtures(input: {
   const wizardAdapter =
     execution.profile.credential === "OPENAI_API_KEY"
       ? execution.profile.id === "runner-codex" ? "paperclip_runner" : "codex_local"
-      : "claude_local";
+      : execution.profile.id === "runner-acpx-claude" ? "paperclip_runner" : "claude_local";
   expect(agents[0].adapterType).toBe(wizardAdapter);
   if (execution.profile.id === "runner-codex") {
     expect(agents[0].adapterConfig?.provider).toBe("codex");
@@ -371,33 +370,9 @@ export async function setupFirstTaskFixtures(input: {
     originalModel: agents[0].adapterConfig?.model ?? null,
   };
   if (execution.profile.id === "runner-acpx-claude") {
-    const runtimePatch = firstTaskNativeRuntimePatch(
-      execution,
-      fixtures,
-      agents[0],
-    );
-    const migrated = await api.patch<Row>(
-      `/api/agents/${agents[0].id}`,
-      runtimePatch,
-    );
-    expect(migrated.adapterType).toBe("paperclip_runner");
-    expect(migrated.adapterConfig?.provider).toBe(execution.profile.provider);
-    expect(migrated.adapterConfig?.instructionsFilePath).toBe(
-      agents[0].adapterConfig?.instructionsFilePath,
-    );
-    expect(migrated.adapterConfig?.paperclipSkillSync?.desiredSkills).toEqual(
-      (
-        runtimePatch.adapterConfig.paperclipSkillSync as {
-          desiredSkills?: unknown[];
-        }
-      )?.desiredSkills,
-    );
-    fixtures.onboardingRuntime = {
-      mode: "post-onboarding-runtime-switch",
-      originalAdapterType: wizardAdapter,
-      testedAdapterType: migrated.adapterType,
-      originalModel: agents[0].adapterConfig?.model ?? null,
-    };
+    expect(agents[0].adapterConfig?.provider).toBe("acpx");
+    expect(agents[0].adapterConfig?.acpxAgent).toBe("claude");
+    expect(agents[0].adapterConfig?.model).toBe("claude-opus-5-5");
   }
   fixtures.agent = {
     id: agents[0].id,

tests/runner-e2e/installed-cli.test.ts

@@ -3,13 +3,17 @@ import { createHash } from "node:crypto";
 import { chmod, lstat, mkdtemp, mkdir, readFile, realpath, rm, symlink, writeFile } from "node:fs/promises";
 import { join } from "node:path";
 import { tmpdir } from "node:os";
-import { afterEach, expect, it } from "vitest";
+import { afterEach, beforeEach, expect, it, vi } from "vitest";
 import { runnerMatrix } from "./catalog.js";
 import { assertInstalledCliSelection, installedCliKeys, probeInstalledCodexStartup, verifyInstalledCli } from "./installed-cli.js";
 import { buildPaperclipServerEnvironment, buildRunnerE2EProcessEnvironment } from "./harness-env.js";
 const cells = runnerMatrix.filter(e => e.profile.id === "runner-acpx-pi");
 const roots: string[] = [];
-afterEach(async () => { await Promise.all(roots.splice(0).map(root => rm(root, { recursive: true, force: true }))); });
+beforeEach(() => {
+  vi.stubEnv("NODE_PATH", undefined);
+  vi.stubEnv("NODE_OPTIONS", undefined);
+});
+afterEach(async () => { vi.unstubAllEnvs(); await Promise.all(roots.splice(0).map(root => rm(root, { recursive: true, force: true }))); });
 const hash = (value: string) => createHash("sha256").update(value).digest("hex");
 async function fixture() {
   const root = await realpath(await mkdtemp(join(tmpdir(), "e2e-installed-cli-"))); roots.push(root);
@@ -187,3 +191,8 @@ it.skipIf(process.platform === "win32")("installed startup rejects stale daemon/
   await expect(probeInstalledCodexStartup({ consumerRoot: wrong.root, serverRoot: wrong.server, expectedRunnerSha256: wrong.expectedRunnerSha256 })).rejects.toThrow("matching native identity");
   expect(await readFile(wrong.observations, "utf8").then(JSON.parse)).toMatchObject({ calls: ["open", "session-close"] });
 });
+
+it("installed startup rejects ambient Node injection before launching", async () => {
+  vi.stubEnv("NODE_PATH", "/untrusted/test-path");
+  await expect(probeInstalledCodexStartup({ consumerRoot: "/unused", serverRoot: "/unused", expectedRunnerSha256: "0".repeat(64) })).rejects.toThrow("forbids NODE_PATH");
+});

tests/runner-e2e/runner.spec.ts

@@ -563,6 +563,9 @@ for (const execution of executions) {
     page,
     request,
   }, testInfo) => {
+    test.skip(execution.suite.id === "first-task" && execution.profile.id === "runner-acpx-claude"
+      && (process.platform !== "darwin" || process.arch !== "arm64"),
+    "Ordinary native Claude onboarding is qualified on Apple Silicon only in this delivery.");
     test.setTimeout(
       execution.task.flow === "warm_three_turn"
         ? deadlineMs