Files
PaperClipAI/tests/runner-e2e/chat-catalog-domains/runtime-computers-privacy.spec.ts
T
f4337bb5fb test(evals): land the chat catalog eval harness on master (#19614)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work
> - The chat-feature evals measure whether agents complete real chat
requests through the product
> - The chat catalog eval harness and its repairs live on the
`chat-eval/master-candidate` branch, not on master
> - Today's work made 139 of 242 chat evals pass in verified single-case
live runs. Each product fix landed on master as its own pull request
> - The eval-side changes (setups, graders, fault seams, evidence
capture) still exist only on the candidate
> - This pull request lands the harness on master. Master stays the
source of truth for product behavior
> - The benefit is one tree: evals run against master commits, and
product fixes reach the evals as soon as they merge

## Linked Issues or Issue Description

Refs #19441, Refs #19599, Refs #20270, Refs #20334, Refs #20342, Refs
#20344, Refs #20354, Refs #20360, Refs #20373, Refs #20382, Refs #20409

**What existing behavior does this improve?**
The chat catalog eval harness and the product it evaluates live on a
long-running branch outside master. Eval fixes do not reach master, so
master-based eval runs use old setups and graders.

**Subsystem affected**
`tests/runner-e2e` chat catalog harness, plus a small set of product
changes that exist only on the candidate.

**Proposed change**
Land the harness on master. Every product fix from this eval work
already merged to master as its own pull request. Conflicts keep
master's product behavior.

## What Changed

- Most of the diff (808 of 851 files) is under `tests/runner-e2e/`:
chat-catalog domains, scenarios, descriptors, the TestOps manifest,
private evidence tooling, and their tests.
- Graders now check saved state, setup literals and rendered or clicked
links. Many keyword and regex checks on free-form replies are removed.
Each changed grader has a bumped `oracleVersion` and calibration tests
(a valid reply, a reworded valid reply, and broken variants that must
fail).
- The shared link helper `hasLink` no longer treats a list link (for
example `/COM/issues?view=mine`) as a link to every task without a URL
key.
- Setup fixes build real state through supported product routes: routine
history, agent grants through the board grant route, budget holds with a
real cap, and a storage-layer archive fault for
`inbox-search.archive-partial-failure`.
- Browser boundary journeys (sign-in, chat-disabled, onboarding) now
write the standard pass evidence files.
- Retargeted cases, approved by the maintainer:
`agent-chat.chat-disabled` (Conference Room gate),
`agents.restore-instructions` (no saved instruction history),
`access.manage-membership` (member permission edits stay board-only).
- The private evidence trusted files (`manual-private-handoff.ts`,
`manual-private-retention.ts`, `manual-private-report-assets.ts`) are
byte-identical to master.
- Product files outside `tests/` (43 files, +1,328 −49) carry
candidate-only changes that master does not yet have. Most are tests.
The main code changes are in `request-human-input.ts`,
`heartbeat-run-summary.ts`, `native-completion-feedback.ts`,
`native-run-finalizer.ts`, `openapi.ts`, `secrets.ts` and
`IssueChatLongThreadPerf.tsx`.

## Verification

- `pnpm vitest run chat-catalog-domains chat-catalog-contract.test.ts
chat-catalog-consistency.test.ts unit-launcher.test.mjs
--testTimeout=180000` in `tests/runner-e2e`: 269 files pass, 3,550 tests
pass.
- `node --test chat-full-catalog-testops-manifest.test.mjs`: 10 pass.
- `node tests/runner-e2e/source-sdk-public-inputs.mjs <HEAD> --check`:
no change.
- `tsc` for `server` and `tests/runner-e2e`: clean.
- Live single-case verification: 139 of 242 cases passed on paid runs,
and a person read the saved evidence for each. These are single-case
runs, not one full-cohort run.

## Risks

- Large diff. Most of it is test-only, but the 43 non-test files change
product code paths. Review those files first.
- Some graders are now less strict on reply wording. A person must read
replies for meaning on those cases.
- 42 cases need disposable outside accounts (Notion, Slack, GitHub,
Daytona and others) and stay blocked. `access.invite-teammate` stays
failing until local agent isolation lands, because the runner's durable
state keeps a live invite token.
- The paid workflow compares the trusted private-evidence files with
master. They are unchanged, so this risk is low.

## Model Used

Anthropic Claude Opus 5.5 (`claude-opus-5-5`), with tool use and code
execution, in Claude Code. Subagents used the same model.

## 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
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-11-10 08:02:11 -06:00

84 lines
11 KiB
TypeScript

/** Opt-in synthetic UI/report sink calibration; no product DB, SSH, provider or native work. */
import { createServer } from "node:http";
import { randomUUID } from "node:crypto";
import { appendFile } from "node:fs/promises";
import { test, expect } from "@playwright/test";
import { createRequire } from "node:module";
import type { WebSocketServer as ServerType } from "ws";
const { WebSocketServer } = createRequire(new URL("../../../server/package.json", import.meta.url))("ws") as { WebSocketServer: typeof ServerType };
import { privateImagePreparation } from "./runtime-computers-image-private.js";
import { privateSshComputer } from "./runtime-computers-private.js";
for (const branch of ["success", "private-child-failure", "parent-failure"] as const) test(`standalone SSH UI privacy ${branch}`, async () => {
test.skip(!process.env.PAPERCLIP_COMPUTER_CANARY_PRIVATE_FILE, "Explicit synthetic-only browser privacy calibration");
const key = `synthetic-ssh-${randomUUID()}`, cookie = `synthetic-cookie-${randomUUID()}`, companyId = randomUUID(), environmentId = randomUUID();
await appendFile(process.env.PAPERCLIP_COMPUTER_CANARY_PRIVATE_FILE!, `${key}\n${cookie}\n`, { mode: 0o600 });
let saved: Record<string, unknown> = {}, keyObserved = false, cookieObserved = false;
const server = createServer(async (req, res) => {
cookieObserved ||= req.headers.cookie === `canary=${cookie}`;
if (req.url?.includes("/company/settings/instance/environments/new")) {
res.setHeader("content-type", "text/html"); res.end(`<form><div><div><label>Driver</label></div><select><option value="ssh">SSH</option></select></div>${["Name", "Host", "Port", "Username", "Remote workspace path", "Private key", "Known hosts"].map(x => `<div><div><label>${x}</label></div>${["Private key", "Known hosts"].includes(x) ? "<textarea></textarea>" : "<input>"}</div>`).join("")}<button>Create environment</button></form><script>document.querySelector('form').onsubmit=async e=>{e.preventDefault();const f=[...document.querySelectorAll('input,textarea')].map(x=>x.value);await fetch('/api/companies/${companyId}/environments',{method:'POST',headers:{'content-type':'application/json'},body:JSON.stringify({name:f[0],config:{host:f[1],port:Number(f[2]),username:f[3],remoteWorkspacePath:f[4],privateKey:f[5],knownHosts:f[6],strictHostKeyChecking:true}})});};</script>`); return;
}
res.setHeader("content-type", "application/json");
if (req.method === "POST") { let raw = ""; for await (const c of req) raw += c; saved = JSON.parse(raw); keyObserved = (saved.config as Record<string, unknown>).privateKey === key;
if (branch === "private-child-failure") { res.statusCode = 500; res.end(JSON.stringify({ diagnostic: key })); return; } res.end(JSON.stringify({ id: environmentId })); return; }
const row = { ...saved, id: environmentId, driver: "ssh", status: "active" }; res.end(JSON.stringify(req.url?.endsWith("/environments") ? [row] : row));
});
await new Promise<void>(resolve => server.listen(0, "127.0.0.1", resolve)); const addr = server.address(); if (!addr || typeof addr === "string") throw new Error("Synthetic server unavailable");
try {
const pending = privateSshComputer({ operation: "create", baseURL: `http://127.0.0.1:${addr.port}`, companyId, companyPrefix: "QA", name: "Development server", host: "synthetic.invalid", port: 22, username: "synthetic", remoteWorkspacePath: "/owned/synthetic", knownHosts: "synthetic-public-host-key", privateKey: key, harness: "claude_local", cookies: [{ name: "canary", value: cookie, domain: "127.0.0.1", path: "/", expires: -1, secure: false, httpOnly: true, sameSite: "Lax" }] });
if (branch === "private-child-failure") await expect(pending).rejects.toThrow("diagnostics withheld"); else { const receipt = await pending; expect(receipt.savedIdentityMatched).toBe(true); expect(receipt.testRunnerAttached).toBe(false); }
expect(keyObserved && cookieObserved).toBe(true); if (branch === "parent-failure") { test.fail(); throw new Error("Deliberate safe synthetic parent failure after private SSH UI"); }
} finally { server.closeAllConnections(); await new Promise<void>(resolve => server.close(() => resolve())); }
});
for (const branch of ["success", "private-terminal-failure", "parent-failure"] as const) test(`standalone image terminal privacy ${branch}`, async () => {
test.skip(!process.env.PAPERCLIP_COMPUTER_CANARY_PRIVATE_FILE, "Explicit synthetic-only terminal privacy calibration");
const token = `synthetic-terminal-${randomUUID()}`, cookie = `synthetic-cookie-${randomUUID()}`, companyId = randomUUID(), environmentId = randomUUID(), sessionId = randomUUID(), templateId = randomUUID();
await appendFile(process.env.PAPERCLIP_COMPUTER_CANARY_PRIVATE_FILE!, `${token}\n${cookie}\n`, { mode: 0o600 });
let authenticated = false, commandObserved = false, promoted = false, cancelled = false;
const session = () => ({ id: sessionId, environmentId, startedByUserId: "synthetic-human", startedByAgentId: null, status: promoted ? "promoted" : "waiting_for_user", promotedTemplateId: promoted ? templateId : null });
const server = createServer(async (req, res) => {
if (req.url === "/") { res.setHeader("content-type", "text/html"); res.end("<html><body>Synthetic operator origin</body></html>"); return; }
for await (const _ of req) { /* synthetic operator requests only */ }
res.setHeader("content-type", "application/json");
if (req.url?.includes("terminal-session-token")) res.end(JSON.stringify({ id: randomUUID(), token, setupSessionId: sessionId, environmentId, websocketPath: "/synthetic-terminal", connectionType: "ssh" }));
else if (req.url?.includes("/finish")) { promoted = true; res.end(JSON.stringify({ session: session(), template: { id: templateId, environmentId, status: "active", templateRef: "synthetic-immutable-provider-id", capturedAt: new Date().toISOString(), createdByUserId: "synthetic-human", createdByAgentId: null } })); }
else if (req.url?.includes("/cancel")) { cancelled = true; res.end("{}"); }
else res.end(JSON.stringify(req.method === "GET" ? session() : { session: session() }));
});
const wss = new WebSocketServer({ server }); wss.on("connection", ws => ws.on("message", raw => {
const frame = JSON.parse(raw.toString());
if (frame.type === "auth") { authenticated = frame.token === token; ws.send(JSON.stringify(branch === "private-terminal-failure" ? { type: "error", message: token } : { type: "ready" })); }
if (frame.type === "input") { commandObserved = /node --version && npm --version && git --version/.test(frame.data); const nonce = /'([a-f0-9]{24})'/.exec(frame.data)?.[1]; ws.send(JSON.stringify({ type: "output", data: `v22.0.0\n10.0.0\ngit version 2.0.0\n__PC_${nonce}_TOOLS_OK__\n` })); }
}));
await new Promise<void>(resolve => server.listen(0, "127.0.0.1", resolve)); const addr = server.address(); if (!addr || typeof addr === "string") throw new Error("Synthetic terminal unavailable");
try {
const pending = privateImagePreparation({ baseURL: `http://127.0.0.1:${addr.port}`, companyId, environmentId, cookies: [{ name: "canary", value: cookie, domain: "127.0.0.1", path: "/", expires: -1, secure: false, httpOnly: true, sameSite: "Lax" }] });
if (branch === "private-terminal-failure") { await expect(pending).rejects.toThrow("diagnostics withheld"); expect(cancelled).toBe(true); }
else { const receipt = await pending; expect(receipt.actualToolsVerified && receipt.promoted && commandObserved).toBe(true); }
expect(authenticated).toBe(true); if (branch === "parent-failure") { test.fail(); throw new Error("Deliberate safe synthetic parent failure after private operator terminal"); }
} finally { for (const ws of wss.clients) ws.terminate(); wss.close(); server.closeAllConnections(); await new Promise<void>(resolve => server.close(() => resolve())); }
});
for (const foreign of [false, true]) test(`saved Morgan runtime binding ${foreign ? "foreign denied" : "exact Codex target"}`, async () => {
test.skip(!process.env.PAPERCLIP_COMPUTER_CANARY_PRIVATE_FILE, "Explicit synthetic-only saved runtime calibration");
const companyId = randomUUID(), environmentId = randomUUID(), agentId = randomUUID(), key = `synthetic-ssh-${randomUUID()}`, cookie = `synthetic-cookie-${randomUUID()}`;
await appendFile(process.env.PAPERCLIP_COMPUTER_CANARY_PRIVATE_FILE!, `${key}\n${cookie}\n`, { mode: 0o600 });
let actualCodexSavedConfig = false;
const config = { host: "synthetic.invalid", port: 22, username: "synthetic", remoteWorkspacePath: "/owned/synthetic", knownHosts: "synthetic-public-key", strictHostKeyChecking: true };
const server = createServer(async (req, res) => {
res.setHeader("content-type", "application/json");
const env = { id: environmentId, name: "Development server", driver: "ssh", status: "active", config };
if (req.method === "GET") { res.end(JSON.stringify(req.url?.includes(`/agents/${agentId}`) ? { id: agentId, companyId: foreign ? randomUUID() : companyId, adapterType: "paperclip_runner", adapterConfig: { provider: "codex", model: "synthetic-saved-model" }, runtimeConfig: { aiConnection: { provider: "openai", mode: "responsible_user" } } } : req.url?.endsWith("/environments") ? [env] : env)); return; }
let raw = ""; for await (const c of req) raw += c; const body = JSON.parse(raw);
if (req.url?.includes("test-environment")) { actualCodexSavedConfig = req.url.includes("/adapters/paperclip_runner/") && body.agentId === agentId && body.environmentId === environmentId && body.adapterConfig.provider === "codex" && body.adapterConfig.model === "synthetic-saved-model" && body.aiConnection.mode === "responsible_user" && body.runner === undefined; res.end(JSON.stringify({ status: "pass", checks: [{ code: "synthetic_saved_runtime" }] })); }
else res.end(JSON.stringify({ ok: true, driver: "ssh", summary: "synthetic transport", details: {} }));
});
await new Promise<void>(resolve => server.listen(0, "127.0.0.1", resolve)); const addr = server.address(); if (!addr || typeof addr === "string") throw new Error("Synthetic runtime unavailable");
try {
const pending = privateSshComputer({ operation: "probe", baseURL: `http://127.0.0.1:${addr.port}`, companyId, companyPrefix: "QA", environmentId, agentId, name: "Development server", ...config, privateKey: key, harness: "saved-agent", cookies: [{ name: "canary", value: cookie, domain: "127.0.0.1", path: "/", expires: -1, secure: false, httpOnly: true, sameSite: "Lax" }] });
if (foreign) { await expect(pending).rejects.toThrow("diagnostics withheld"); expect(actualCodexSavedConfig).toBe(false); }
else { const r = await pending; expect(r.agentId === agentId && r.selectedHarness === "codex_local" && actualCodexSavedConfig).toBe(true); }
} finally { server.closeAllConnections(); await new Promise<void>(resolve => server.close(() => resolve())); }
});