Files
paperclip/scripts/runner-api-eval-worker.ts
Devin FoleyandPaperclip be6f49a425 feat(runner): refresh shared coding harness runtimes (#13838)
## Thinking Path

> - Paperclip runs agents through local adapters and the native runner.
> - Both paths must use the same installed provider CLI.
> - New models require current harness releases.
> - The runner still pins Codex 0.153.4, Claude SDK 0.3.263, and
OpenCode 1.18.29.
> - Changing the image alone would fail the runner's exact version and
executable checks.
> - This pull request updates those dependencies, integrity checks,
controller checks, and image pins together.
> - Shared installations can then run the current models without a
task-time download.

## Linked Issues or Issue Description

Refs #13829, which updates model choices and reasoning controls.
Searches found no open PR that updates these runtime pins.

**Current behavior**

The shared provider pack ships old CLIs. Claude Code 2.1.263 cannot run
Opus 5.5, which requires 2.1.280. Remote controllers reject provider
packs whose versions differ from their declared pins.

**Proposed behavior**

Use Codex 0.156.0, Claude Agent SDK 0.3.280 / Claude Code 2.1.280, and
OpenCode 1.18.32 throughout the runner. Keep the reviewed ACP bridge
patches and one shared CLI installation per provider.

**Reason and benefit**

Current harnesses support the new model IDs while preserving executable
verification and remote provider-pack compatibility checks.

## What Changed

- Update dependency overrides, the Codex ACP package patch, runtime
profiles, and remote controller pins.
- Verify the new Claude Linux x64 and macOS arm64/x64 executables and
Codex Linux x64 executable against integrity-verified npm archives.
- Refresh OpenCode version checks, fixtures, and the runner
configuration label.
- Refresh the eval image's Grok, Gemini, Kimi, Cursor, and GitHub CLI
pins and archive hashes. Hermes remains current at 0.19.0.
- Refresh the build-time lock digest from clean pnpm 9.15.4 resolution.
Leave lockfile commits to repository automation.
- Document model compatibility and the separation between CLI runtimes
and patched ACP bridges.

## Verification

- `pnpm -r typecheck` and `pnpm build` passed.
- Rust workspace release tests passed.
- Package/patch and OpenCode binary-materialization contract tests: 11
passed.
- Real Codex 0.156.0 startup-ownership and paginated session-resume
probes passed with isolated synthetic homes and no model turn.
- Codex app-server `thread/start` preserved `gpt-6-sol` and
`gpt-6-luna`; no `turn/start` was sent. An unauthenticated built-in
catalog does not include those account-served entries.
- Installed Claude integrity probes passed for `claude-opus-5-5` and
`claude-fable-5-1`.
- `pnpm --filter @paperclipai/paperclip-runner
test:opencode:qualification` passed with the actual OpenCode 1.18.32
executable under Node 24 and Node 25. The loopback provider exercise
covers health/version, session creation/read/delete, SSE, and a
completed async prompt.
- `pnpm check:token-gates` passed.
- The targeted runner suite passed 130 tests. Three macOS failures in
snapshot module lookup and OpenCode final-message selection also
reproduce on the unchanged base; Linux CI will provide the platform
check.
- [Final Linux
CI](https://github.com/paperclipai/paperclip/actions/runs/35798076399):
all gates passed. Four jobs needed one retry after their CI workers
received shutdown signals. The PR has 55 successful checks, two skipped
checks, Greptile 5/5, and no unresolved review threads.
- Changed runner configuration UI tests: 5 passed.
- Full macOS `pnpm test:run` reached 13,094 passing server tests, 84
skipped, and 18 failures before the wrapper stopped. Failures involved
skill-cache publication permissions, missing bundled connector skills in
the worktree, and a conversation-reset timing case. The 10 cache
permission failures reproduce on the unchanged base; both
conversation-reset cases passed on a targeted retry. The wrapper did not
reach its later workspace/serialized groups locally; Linux CI covers
those groups.
- The local Docker daemon did not respond, so no local Docker build was
run. No billable model requests were made.

## Risks

- Deploy the matching controller and provider pack together. Older
controllers enforce their previous exact pins.
- Current upstream CLIs can change behavior. Existing protocol tests and
isolated real Codex probes cover the integration boundaries;
authenticated model inference is not part of these checks.
- ACP bridge package versions and executable digests stay unchanged
because their executable bytes are unchanged. Only the underlying
CLI/SDK dependencies move.
- No schema migration. Revert the runtime and image pins together to
roll back.

## Model Used

OpenAI GPT-6 via Codex, with repository tools, code execution, and web
research. The exact serving model ID and context window were not exposed
by 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
- [x] I have run tests locally and they pass for the changed surfaces
and real-executable probes; full macOS-suite limitations are listed
above
- [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>
2026-09-22 17:02:29 -07:00

223 lines
18 KiB
TypeScript
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
import { AGENT_CHAT_DIRECTIVE } from "../server/src/services/agent-conversations.js";
/** JSONL worker for the companion paperclip-evals API suite. Never selects cases or retries. */
import { createHash, randomUUID } from "node:crypto";
import { createReadStream, realpathSync } from "node:fs";
import { execFileSync } from "node:child_process";
import { mkdir, writeFile, readFile } from "node:fs/promises";
import { join, resolve } from "node:path";
import { createInterface } from "node:readline";
import { homedir } from "node:os";
import { runnerApiCatalog } from "../server/src/services/native-runtime/runner-api-catalog.js";
import { startRunnerApiTestServer } from "../server/src/__tests__/helpers/runner-api-server.js";
import { registerRunnerPrpAuthority } from "../server/src/realtime/runner-prp-ws.js";
import { createRunnerdCodexTransport, defaultCapabilityRunnerdBinary } from "../packages/paperclip-runner/src/live/runnerd-codex-transport.js";
import { createSkilllessCodexThreadConfig } from "../packages/paperclip-runner/src/drivers/codex/codex-app-server-driver.js";
import { AttemptJournal } from "../packages/paperclip-runner/src/evals/attempt-journal.js";
import { estimateModelCostNanodollars } from "../packages/paperclip-runner/src/evals/model-pricing.js";
import { CONNECTION_INTENT_AGENT_GUIDANCE } from "../packages/shared/src/connection-intent-guidance.js";
if (process.argv.includes("--catalog")) {
process.stdout.write(JSON.stringify(runnerApiCatalog()) + "\n");
process.exit(0);
}
if (!process.argv.includes("--jsonl")) throw new Error("Use --catalog or --jsonl; there is no default campaign");
process.env.PAPERCLIP_AGENT_JWT_SECRET = randomUUID() + randomUUID();
const output = (value: unknown) => process.stdout.write("RUNNER_API_EVAL " + JSON.stringify(value) + "\n");
const OPENROUTER_MODELS = new Set(["openrouter/anthropic/claude-sonnet-5", "openrouter/deepseek/deepseek-v4-flash-0731", "openrouter/google/gemini-3.8-flash"]);
// The controller selects and injects one provider credential. The worker never
// reads ambient home credential files or desktop keychains.
function openRouterEnvironment() {
const key = process.env.OPENROUTER_API_KEY;
if (!key || /\s/.test(key)) throw new Error("Controller must inject OPENROUTER_API_KEY");
return { PATH: process.env.PATH, OPENROUTER_API_KEY: key };
}
const server = await startRunnerApiTestServer();
const fileDigests = new Map<string, Promise<string | null>>();
function fileDigest(path: string): Promise<string | null> {
let digest = fileDigests.get(path);
if (!digest) {
digest = (async () => {
const hash = createHash("sha256");
for await (const chunk of createReadStream(path)) hash.update(chunk);
return hash.digest("hex");
})().catch(() => null);
fileDigests.set(path, digest);
}
return digest;
}
const runtimeBuild = {
runnerBinarySha256: await fileDigest(defaultCapabilityRunnerdBinary()),
workerSourceSha256: await fileDigest(resolve("scripts/runner-api-eval-worker.ts")),
lockfileSha256: await fileDigest(resolve("pnpm-lock.yaml")),
};
output({ ready: true });
const record = (value: unknown): Record<string, any> => value && typeof value === "object" ? value as Record<string, any> : {};
try {
for await (const line of createInterface({ input: process.stdin })) {
const request = JSON.parse(line);
if (request.shutdown) break;
const directory = resolve(request.outputDirectory);
await mkdir(directory, { recursive: true, mode: 0o700 });
const journal = new AttemptJournal(join(directory, "journal.jsonl"));
const startedAt = new Date().toISOString();
journal.append({ kind: "attempt_started", attemptId: request.attemptId, model: request.model, startedAt });
const started = performance.now();
const isOpenRouter = OPENROUTER_MODELS.has(request.model);
const provider = isOpenRouter ? "opencode" : request.model === "claude-sonnet-5" ? "acpx" : "codex";
let providerVersion: string | null = null;
const fixture = await server.fixture({ mode: request.mode, apiToolsEnabled: request.arm !== "baseline", reset: true, conversation: request.conversation === true, connectionScenario: request.connectionScenario });
const initialState = await fixture.snapshot();
const substitutions = Object.fromEntries(Object.entries(fixture).filter(([, value]) => typeof value === "string"));
const expand = (value: any): any => typeof value === "string" ? value.replace(/\{\{(\w+)\}\}/g, (_, key) => String(substitutions[key] ?? (() => { throw new Error(`Unknown fixture variable ${key}`); })())) : Array.isArray(value) ? value.map(expand) : value && typeof value === "object" ? Object.fromEntries(Object.entries(value).map(([key, entry]) => [key, expand(entry)])) : value;
const calls: Record<string, any>[] = [], notifications: unknown[] = [];
const observedToolCallIds = new Set<string>();
const definitions = await fixture.authority.definitions();
const prompt = expand(request.prompt ?? "");
let error: string | null = null, evidence: unknown = null, thread: Record<string, any> = {}, usage: Record<string, any> | null = null;
let usageUpdates = 0, lastUsageFingerprint = "", terminalSeen = false, providerTurnStarted = false;
let diagnosticTail: string | null = null;
let bundle: ReturnType<typeof createRunnerdCodexTransport> | undefined;
let timer: ReturnType<typeof setTimeout> | undefined;
const invoke = async (tool: string, args: unknown, callId: string) => {
observedToolCallIds.add(callId);
if (observedToolCallIds.size > 12) throw new Error("Attempt tool-call ceiling reached");
const call: Record<string, any> = { tool, arguments: args, callId, startedMs: performance.now() - started };
calls.push(call);
journal.append({ kind: "tool_dispatch", ...call });
try { call.result = await fixture.authority.execute({ tool, arguments: args, callId }); }
catch (caught) { call.error = caught instanceof Error ? caught.message : String(caught); call.status = record(caught).status; }
call.durationMs = performance.now() - started - call.startedMs;
journal.append({ kind: "tool_result", ...call });
return call;
};
try {
if (request.calls) {
for (const [i, call] of request.calls.entries()) await invoke(call.tool, expand(call.arguments), call.callId ?? `direct-${i}`);
usage = { inputTokens: 0, outputTokens: 0, cachedInputTokens: 0, estimatedCostNanodollars: 0, providerRequests: 0, accountingProvenance: "Provider-free direct authority contract; no provider or runnerd dispatched" };
} else {
if (request.model === "claude-sonnet-5" && (process.platform !== "linux" || process.arch !== "x64")) throw new Error("Qualified ACPX Claude requires Linux x64; no provider turn was dispatched");
if (!request.reservationId || request.maxCostUsd !== 0.5 || !["gpt-5.6-luna", "claude-sonnet-5", ...OPENROUTER_MODELS].includes(request.model)) throw new Error("Paid attempt requires ledger reservation and qualified model");
if (isOpenRouter) {
providerVersion = execFileSync(resolve("packages/paperclip-runner/node_modules/opencode-ai/bin/opencode.exe"), ["--version"], { encoding: "utf8" }).trim();
if (providerVersion !== "1.18.32") throw new Error("OpenCode profile requires version 1.18.32");
}
const providerEnvironment = isOpenRouter ? openRouterEnvironment() : request.model === "claude-sonnet-5" ? (() => {
const token = process.env.CLAUDE_CODE_OAUTH_TOKEN;
if (!token) throw new Error("Controller must inject CLAUDE_CODE_OAUTH_TOKEN");
return { PATH: process.env.PATH, CLAUDE_CODE_OAUTH_TOKEN: token };
})() : undefined;
bundle = createRunnerdCodexTransport({
provider, acpxAgent: "claude", acpxPermissionMode: "approve-reads",
environment: { ...providerEnvironment, PAPERCLIP_PROVIDER_TRACE_PATH: join(directory, "provider-trace.jsonl"), PAPERCLIP_PROVIDER_TRACE_MAX_BYTES: String(32 * 1024 * 1024) },
codexCommand: request.model === "gpt-5.6-luna" ? realpathSync(execFileSync("which", ["codex"], { encoding: "utf8" }).trim()) : undefined,
sourceCodexHome: process.env.CODEX_HOME ?? join(homedir(), ".codex"),
runnerBinary: defaultCapabilityRunnerdBinary(), stateDirectory: join(server.root, `runner-${request.attemptId}`),
lifecyclePolicy: { mode: "per_turn", idleTimeoutMs: null },
prpIdentity: { runnerInstanceId: `runner-${fixture.runId}`, environmentLeaseId: `lease-${fixture.runId}`, runId: fixture.runId, normalizedSessionId: `session-${fixture.runId}`, turnId: `turn-${fixture.runId}`, itemId: `item-${fixture.runId}` },
controlPlaneRegistration: (prp) => registerRunnerPrpAuthority({ companyId: fixture.companyId, runId: fixture.runId, authority: prp }),
});
const active = bundle;
timer = setTimeout(() => { error = "Attempt exceeded 120 seconds"; void active.transport.close(error); }, 120_000);
bundle.transport.setServerRequestHandler(async (request) => {
const params = record(request.params);
const result = await invoke(String(params.tool), params.arguments, String(params.callId));
return { success: !result.error, contentItems: [{ type: "inputText", text: JSON.stringify(result.error ? { ok: false, error: result.error, status: result.status } : { ok: true, result: result.result }) }] };
});
await bundle.transport.request("initialize", {});
thread = await bundle.transport.request("thread/start", {
cwd: fixture.workspace, model: request.model,
completionContract: { revision: "runner-api-eval-v1", criterionIds: ["objective"] },
config: { ...createSkilllessCodexThreadConfig(fixture.workspace), model_reasoning_effort: "low" },
permissions: "paperclip-runner-workspace-only", runtimeWorkspaceRoots: [fixture.workspace], approvalPolicy: "never",
baseInstructions: "You are operating a disposable real Paperclip company. Use the provided tools to do the user's task. Do not use shell, network, skills, or credentials. Stop when the requested work is verified. " + (request.arm === "baseline" ? "" : "Prefer available dedicated tools. Only use search_api and call_api when no dedicated tool supports the required operation or parameters. Do not search before ordinary dedicated tool use.") + "\n" + CONNECTION_INTENT_AGENT_GUIDANCE + (request.conversation ? "\n" + AGENT_CHAT_DIRECTIVE : ""),
dynamicTools: definitions, experimentalRawEvents: true, persistExtendedHistory: true,
});
if (request.preflight) {
terminalSeen = true;
usage = { inputTokens: 0, outputTokens: 0, cachedInputTokens: 0, estimatedCostNanodollars: 0, providerRequests: 0, preflight: true };
} else {
journal.append({ kind: "provider_turn_dispatch", thread, at: new Date().toISOString() });
providerTurnStarted = true;
await bundle.transport.request("turn/start", { input: [{ type: "text", text: prompt }] });
for await (const notification of bundle.transport.notifications()) {
const retained = { ...notification, observedAt: new Date().toISOString() };
journal.append({ kind: "notification", ...retained });
notifications.push(retained);
const item = record(record(notification.params).item);
if (["dynamicToolCall", "tool_call", "tool_result"].includes(item.type) && typeof item.id === "string") {
observedToolCallIds.add(item.id);
if (observedToolCallIds.size > 12) { error = "Attempt tool-call ceiling reached, including provider-rejected calls"; break; }
}
if (notification.method === "thread/tokenUsage/updated") {
const params = record(notification.params);
const tokenUsage = record(params.tokenUsage);
const cumulative = record(tokenUsage.total ?? tokenUsage.totalTokenUsage ?? tokenUsage.total_token_usage ?? tokenUsage);
if (request.model === "claude-sonnet-5" && params.runDeltaAvailable !== true) continue;
const total = request.model === "claude-sonnet-5" ? record(tokenUsage.runDelta) : cumulative;
const cacheWriteTokens = total.cacheWriteTokens ?? 0;
const inputTokens = (total.inputTokens ?? total.input_tokens) + (request.model === "claude-sonnet-5" ? (total.cacheReadTokens ?? 0) : isOpenRouter ? (total.cacheReadTokens ?? total.cachedInputTokens ?? 0) : 0);
const outputTokens = total.outputTokens ?? total.output_tokens;
const cachedInputTokens = total.cacheReadTokens ?? total.cachedInputTokens ?? total.cache_read_input_tokens ?? 0;
if ([inputTokens, outputTokens, cachedInputTokens, cacheWriteTokens].every(value => Number.isSafeInteger(value) && value >= 0)) {
const fingerprint = `${inputTokens}:${outputTokens}:${cachedInputTokens}`;
if (fingerprint !== lastUsageFingerprint) { usageUpdates++; lastUsageFingerprint = fingerprint; }
usage = { inputTokens, outputTokens, cachedInputTokens, cacheWriteTokens, providerRequests: (total.requests ?? total.providerRequests ?? 0) > 0 ? total.requests ?? total.providerRequests : null, providerUsageUpdateCount: usageUpdates, providerRequestCountSource: (total.requests ?? total.providerRequests ?? 0) > 0 ? "provider" : "not reported; providerUsageUpdateCount is a proxy", providerReportedCostNanodollars: cumulative.providerCostUsd > 0 ? Math.round(cumulative.providerCostUsd * 1e9) : null, ...estimateModelCostNanodollars(request.model, { inputTokens, outputTokens, cachedInputTokens }) };
// A conservative one-hour cache-write ceiling (2× standard input)
// covers either cache duration when the runtime omits its TTL.
if (request.model === "claude-sonnet-5" || request.model === "openrouter/anthropic/claude-sonnet-5") {
usage.estimatedCostNanodollars += cacheWriteTokens * 4_000;
usage.cacheWriteUsdPerMillionTokens = 4;
usage.cacheWritePricing = "Conservative one-hour cache-write rate; TTL not exposed";
}
if (Math.max(usage.estimatedCostNanodollars, usage.providerReportedCostNanodollars ?? 0) >= 500_000_000) { error = "Attempt cost ceiling reached"; break; }
}
}
if (notification.method === "turn/completed") {
terminalSeen = true;
const params = record(notification.params);
const terminal = record(params.turn);
if ((params.status ?? terminal.status) === "failed") error = String(record(params.error ?? terminal.error).message ?? "Provider turn failed");
break;
}
}
}
evidence = bundle.evidence();
if (!request.preflight && (!usage || usage.inputTokens + usage.outputTokens === 0)) {
usage = null;
error ??= "Missing provider accounting; block further paid attempts";
}
}
} catch (caught) { error ??= caught instanceof Error ? caught.message : String(caught); }
finally {
if (timer) clearTimeout(timer);
if (bundle) {
evidence = bundle.evidence();
diagnosticTail = await readFile(join(server.root, `runner-${request.attemptId}`, "diagnostics", "runnerd.stderr.log"), "utf8").then(text => text.slice(-16000).replace(/sk-(?:ant-|or-v1-)[A-Za-z0-9_-]+/g, "[redacted]").replace(/Bearer\s+[^\s"']+/gi, "Bearer [redacted]").replace(/eyJ[A-Za-z0-9_-]+\.[A-Za-z0-9_-]+\.[A-Za-z0-9_-]+/g, "[redacted JWT]")).catch(() => null);
}
await bundle?.transport.close();
}
if (!providerTurnStarted && !request.calls) {
usage = { inputTokens: 0, outputTokens: 0, cachedInputTokens: 0, estimatedCostNanodollars: 0, providerRequests: 0, accountingProvenance: "No provider turn/start was dispatched" };
}
const artifact = {
schema: "paperclip-runner/eval-session-artifact/v1", attemptId: request.attemptId,
requestedModel: request.calls ? "provider-free" : request.model ?? "provider-free", provider: request.calls ? "none" : provider, driver: request.calls ? "direct-authority-contract" : "real-server-api-tools",
evidenceMode: request.calls ? "provider-free-contract" : "live-provider",
providerSessionId: record(thread.thread).id ?? null, effectiveModel: record(thread.thread).model ?? null,
providerVersion: request.calls ? null : request.model === "gpt-5.6-luna" ? execFileSync("codex", ["--version"], { encoding: "utf8" }).trim() : providerVersion ?? record(evidence).providerVersion ?? null,
runtimeVersions: { node: process.versions.node, acpx: record(evidence).providerVersion, agentServer: record(evidence).agentServerVersion, agentRuntime: record(evidence).agentRuntimeVersion },
runtimeBuild,
timing: { startedAt, finishedAt: new Date().toISOString(), durationMs: performance.now() - started },
usage, accountingComplete: Boolean(request.calls) || !providerTurnStarted || terminalSeen, providerTurnStarted, observedProviderToolCalls: observedToolCallIds.size, diagnosticTail, error, calls, notifications, evidence, thread, prompt, arm: request.arm ?? "treatment",
fixture: substitutions, initialState, state: await fixture.snapshot(),
toolSchemaBytes: Buffer.byteLength(JSON.stringify(definitions)), definitions,
};
// The controller supplies a unique retained attempt directory; never overwrite evidence.
journal.append({ kind: "attempt_finished", error, usage, terminalSeen, at: new Date().toISOString() });
journal.close();
await writeFile(join(directory, "artifact.json"), JSON.stringify(artifact, null, 2), { flag: "wx", mode: 0o600 });
output({ attemptId: request.attemptId, artifactPath: join(directory, "artifact.json"), error, usage });
}
} finally { await server.close(); }