mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-13 15:55:46 +01:00
## 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>
51 lines
6.1 KiB
TypeScript
51 lines
6.1 KiB
TypeScript
import { readFile } from "node:fs/promises";
|
|
import { randomBytes, createHash } from "node:crypto";
|
|
|
|
import { chromium, test } from "@playwright/test";
|
|
import type { ImagePreparationRequest, ImagePreparationReceipt } from "./runtime-computers-image-private.js";
|
|
import type { EnvironmentCustomImageSetupSession, EnvironmentCustomImageTemplate, EnvironmentCustomImageTerminalSessionToken } from "../../../packages/shared/src/validators/environment-custom-images.js";
|
|
let phase: "input" | "session" | "terminal-token" | "browser" | "terminal" | "capture" | "saved-proof" = "input";
|
|
async function main(): Promise<ImagePreparationReceipt> {
|
|
let attached = false; try { test.info(); attached = true; } catch { /* standalone operator child */ } if (attached) throw new Error("Image terminal cannot run inside Test");
|
|
const r = JSON.parse(await readFile(process.argv[2]!, "utf8")) as ImagePreparationRequest & { executablePath: string }, base = new URL(r.baseURL);
|
|
if (base.protocol !== "http:" || base.hostname !== "127.0.0.1" || ![r.companyId, r.environmentId].every(x => /^[a-f0-9-]{36}$/i.test(x))) throw new Error("Invalid owned image binding");
|
|
const headers = { Cookie: r.cookies.map(x => `${x.name}=${x.value}`).join("; "), Origin: base.origin, "Content-Type": "application/json" };
|
|
async function api<T>(pathname: string, body?: unknown) {
|
|
const response = await fetch(new URL(pathname, base), { method: body === undefined ? "GET" : "POST", headers, body: body === undefined ? undefined : JSON.stringify(body), redirect: "error", signal: AbortSignal.timeout(300000) });
|
|
if (!response.ok) { await response.body?.cancel(); throw new Error("Image operator public action rejected"); } return response.json() as Promise<T>;
|
|
}
|
|
let sessionId: string | undefined, promoted = false;
|
|
try {
|
|
phase = "session"; const start = await api<{ session: EnvironmentCustomImageSetupSession }>(`/api/environments/${r.environmentId}/custom-image-setup-sessions?companyId=${r.companyId}`, { ttlSeconds: 600 }); sessionId = start.session.id;
|
|
if (start.session.environmentId !== r.environmentId || !start.session.startedByUserId || start.session.startedByAgentId || start.session.status !== "waiting_for_user") throw new Error("Actual owned human setup session is not interactive");
|
|
phase = "terminal-token"; const token = await api<EnvironmentCustomImageTerminalSessionToken>(`/api/environment-custom-image-setup-sessions/${sessionId}/terminal-session-token?companyId=${r.companyId}`, {});
|
|
const url = new URL(token.websocketPath, base); if (url.origin !== base.origin || token.setupSessionId !== sessionId || token.environmentId !== r.environmentId) throw new Error("Terminal token is not the exact setup session"); url.protocol = "ws:";
|
|
const nonce = randomBytes(12).toString("hex"), marker = `__PC_${nonce}_TOOLS_OK__`;
|
|
phase = "browser"; const browser = await chromium.launch({ headless: true, executablePath: r.executablePath });
|
|
try {
|
|
const context = await browser.newContext(); await context.addCookies(r.cookies); const page = await context.newPage(); await page.goto(base.origin);
|
|
phase = "terminal";
|
|
// A literal browser script avoids TS-loader helper references crossing CDP.
|
|
const script = String.raw`(({ url, token, nonce, marker }) => new Promise((resolve, reject) => {
|
|
const socket = new WebSocket(url); let output = "", settled = false;
|
|
const timeout = setTimeout(() => finish(new Error("Actual image tools timed out")), 120000);
|
|
function finish(error) { if (settled) return; settled = true; clearTimeout(timeout); socket.close(); error ? reject(error) : resolve(); }
|
|
socket.onopen = () => socket.send(JSON.stringify({ type: "auth", token }));
|
|
socket.onmessage = event => { let frame; try { frame = JSON.parse(String(event.data)); } catch { return; }
|
|
if (frame.type === "ready") socket.send(JSON.stringify({ type: "input", data: "(node --version && npm --version && git --version) && printf '\\n__PC_%s_TOOLS_OK__\\n' '" + nonce + "'\r" }));
|
|
if (frame.type === "error") finish(new Error("Actual setup terminal rejected"));
|
|
if (frame.type === "output" && typeof frame.data === "string") { output += frame.data; if (output.length > 262144) return finish(new Error("Private output bound exceeded")); if (output.includes(marker) && /v\d+\.\d+\.\d+/.test(output) && /git version \d+\./.test(output)) finish(); }
|
|
}; socket.onerror = () => finish(new Error("Actual setup terminal disconnected")); socket.onclose = () => { if (!settled) finish(new Error("Actual setup terminal closed early")); };
|
|
}))`;
|
|
await page.evaluate(`${script}(${JSON.stringify({ url: url.href, token: token.token, nonce, marker })})`);
|
|
} finally { await browser.close(); }
|
|
phase = "capture"; const finish = await api<{ session: EnvironmentCustomImageSetupSession; template: EnvironmentCustomImageTemplate }>(`/api/environment-custom-image-setup-sessions/${sessionId}/finish?companyId=${r.companyId}`, {});
|
|
phase = "saved-proof"; const saved = await api<EnvironmentCustomImageSetupSession>(`/api/environment-custom-image-setup-sessions/${sessionId}?companyId=${r.companyId}`);
|
|
if (finish.template.environmentId !== r.environmentId || finish.template.status !== "active" || !finish.template.templateRef || !finish.template.capturedAt || !finish.template.createdByUserId || finish.template.createdByAgentId || saved.status !== "promoted" || saved.promotedTemplateId !== finish.template.id) throw new Error("Actual provider image was not captured/promoted"); promoted = true;
|
|
return { processIsolation: true, testRunnerAttached: false, sessionId, templateId: finish.template.id, templateRefSha256: createHash("sha256").update(finish.template.templateRef).digest("hex"), actualToolsVerified: true, promoted: true, sourceEnvironmentId: r.environmentId };
|
|
} finally {
|
|
if (sessionId && !promoted) await api(`/api/environment-custom-image-setup-sessions/${sessionId}/cancel?companyId=${r.companyId}`, { reason: "Owned eval operator setup cleanup" });
|
|
}
|
|
}
|
|
main().then(x => process.stdout.write(JSON.stringify(x))).catch(() => { process.stderr.write(`image-private-phase:${phase}`); process.exitCode = 1; });
|