mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-14 00:14:02 +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>
129 lines
8.4 KiB
TypeScript
129 lines
8.4 KiB
TypeScript
import { afterEach, describe, expect, it, vi } from "vitest";
|
|
import type { APIRequestContext, APIResponse } from "@playwright/test";
|
|
import { pollUntil, RunnerApi, RunnerApiHttpError } from "./api.js";
|
|
const response = (status: number, data: unknown = {}): APIResponse => ({
|
|
ok: () => status >= 200 && status < 300, status: () => status,
|
|
json: async () => data, text: async () => JSON.stringify(data), url: () => "http://fixture.invalid/instructions",
|
|
}) as APIResponse;
|
|
afterEach(() => { vi.unstubAllEnvs(); vi.unstubAllGlobals(); });
|
|
describe("fixture instruction revision fence", () => {
|
|
function api(current: APIResponse, saved = response(200)) {
|
|
vi.stubEnv("PAPERCLIP_RUNNER_E2E_PORT", "3100");
|
|
const request = { get: vi.fn(async () => current), put: vi.fn(async () => saved) };
|
|
return { request, api: new RunnerApi(request as unknown as APIRequestContext) };
|
|
}
|
|
it.each([[404, {}, null], [200, { contentHash: "current-hash" }, "current-hash"]])("saves from a %s read with its exact base", async (status, detail, baseHash) => {
|
|
const fixture = api(response(status as number, detail));
|
|
await fixture.api.saveAgentInstructions("agent", "Fixture instructions");
|
|
expect(fixture.request.get).toHaveBeenCalledWith("/api/agents/agent/instructions-bundle/file?path=AGENTS.md");
|
|
expect(fixture.request.put).toHaveBeenCalledExactlyOnceWith("/api/agents/agent/instructions-bundle/file", {
|
|
data: { path: "AGENTS.md", content: "Fixture instructions", baseHash },
|
|
});
|
|
});
|
|
it.each([response(403), response(500), response(200, {})])("never treats a failed or incomplete read as a new file", async current => {
|
|
const fixture = api(current);
|
|
await expect(fixture.api.saveAgentInstructions("agent", "Fixture instructions")).rejects.toThrow();
|
|
expect(fixture.request.put).not.toHaveBeenCalled();
|
|
});
|
|
it("surfaces a concurrent edit without retrying or overwriting it", async () => {
|
|
const fixture = api(response(200, { contentHash: "old" }), response(409, { error: "Revision conflict" }));
|
|
await expect(fixture.api.saveAgentInstructions("agent", "Fixture instructions")).rejects.toThrow("409");
|
|
expect(fixture.request.get).toHaveBeenCalledTimes(1);
|
|
expect(fixture.request.put).toHaveBeenCalledTimes(1);
|
|
});
|
|
});
|
|
|
|
|
|
it("passes a bounded GET timeout to the request transport without changing its response", async () => {
|
|
vi.stubEnv("PAPERCLIP_RUNNER_E2E_PORT", "3100");
|
|
const get = vi.fn(async () => response(200, { status: "running" }));
|
|
const api = new RunnerApi({ get } as unknown as APIRequestContext);
|
|
await expect(api.get("/api/heartbeat-runs/run", { timeout: 1234 })).resolves.toEqual({ status: "running" });
|
|
expect(get).toHaveBeenCalledExactlyOnceWith("/api/heartbeat-runs/run", { timeout: 1234 });
|
|
});
|
|
|
|
|
|
it("preserves GET HTTP status separately from diagnostic text", async () => {
|
|
vi.stubEnv("PAPERCLIP_RUNNER_E2E_PORT", "3100");
|
|
const api = new RunnerApi({ get: vi.fn(async () => response(503, { error: "PRIVATE" })) } as unknown as APIRequestContext);
|
|
const error = await api.get("/api/environments/env/leases").catch(error => error);
|
|
expect(error).toBeInstanceOf(RunnerApiHttpError);
|
|
if (!(error instanceof RunnerApiHttpError)) throw new Error("Expected typed HTTP error");
|
|
expect(error.status).toBe(503);
|
|
});
|
|
|
|
|
|
it("a strict observation cannot pass after first observing invalid evidence", async () => {
|
|
const load = vi.fn().mockRejectedValueOnce(new Error("invalid evidence")).mockResolvedValue(true);
|
|
await expect(pollUntil({ label: "strict proof", deadlineAt: Date.now() + 1000,
|
|
load, accept: Boolean, intervalMs: 1, retryOnError: () => false })).rejects.toThrow("invalid evidence");
|
|
expect(load).toHaveBeenCalledTimes(1);
|
|
});
|
|
|
|
|
|
it("keeps existing polling retries when the strict error policy is absent", async () => {
|
|
const load = vi.fn().mockRejectedValueOnce(new Error("temporary transport failure")).mockResolvedValue(true);
|
|
await expect(pollUntil({ label: "ordinary state", deadlineAt: Date.now() + 1000,
|
|
load, accept: Boolean, intervalMs: 1 })).resolves.toBe(true);
|
|
expect(load).toHaveBeenCalledTimes(2);
|
|
});
|
|
|
|
it("observes identical native poll results with the actual actor and no traced request", async () => {
|
|
vi.stubEnv("PAPERCLIP_RUNNER_E2E_PORT", "3100");
|
|
const get = vi.fn(() => { throw new Error("Poll must not enter the Playwright API recorder"); });
|
|
const api = new RunnerApi({ get } as unknown as APIRequestContext);
|
|
api.setBrowserSession("synthetic-cookie");
|
|
const rows = [{ id: "11111111-1111-4111-8111-111111111111", status: "running" }];
|
|
const fetcher = vi.fn(async (_input: RequestInfo | URL, _init?: RequestInit) => new Response(JSON.stringify(rows), { status: 200 }));
|
|
vi.stubGlobal("fetch", fetcher);
|
|
const endpoint = "/api/companies/22222222-2222-4222-8222-222222222222/heartbeat-runs?limit=100";
|
|
await expect(api.getRunObservation(endpoint, { timeout: 1234 })).resolves.toEqual(rows);
|
|
expect(get).not.toHaveBeenCalled(); expect(fetcher).toHaveBeenCalledTimes(1);
|
|
expect(String(fetcher.mock.calls[0]![0])).toBe(`http://127.0.0.1:3100${endpoint}`);
|
|
expect(fetcher.mock.calls[0]![1]).toMatchObject({ method: "GET", redirect: "error", headers: { Cookie: "synthetic-cookie", Origin: "http://127.0.0.1:3100" } });
|
|
expect((fetcher.mock.calls[0]![1] as RequestInit).signal).toBeInstanceOf(AbortSignal);
|
|
});
|
|
it.each([
|
|
"/api/companies/22222222-2222-4222-8222-222222222222/issues?includeAgentCreatedTasks=true&limit=1000",
|
|
"/api/companies/22222222-2222-4222-8222-222222222222/activity?limit=1000",
|
|
])("reads the repeated catalog inventory observation %s outside the Playwright trace", async endpoint => {
|
|
vi.stubEnv("PAPERCLIP_RUNNER_E2E_PORT", "3100");
|
|
const get = vi.fn(() => { throw new Error("Inventory poll must not enter the Playwright API recorder"); });
|
|
const fetcher = vi.fn(async () => new Response("[]", { status: 200 })); vi.stubGlobal("fetch", fetcher);
|
|
await expect(new RunnerApi({ get } as unknown as APIRequestContext).getRunObservation(endpoint)).resolves.toEqual([]);
|
|
expect(get).not.toHaveBeenCalled(); expect(String((fetcher.mock.calls[0] as unknown[])[0])).toBe(`http://127.0.0.1:3100${endpoint}`);
|
|
});
|
|
it("keeps local-trusted run observations cookie-free and preserves typed HTTP failure without retry", async () => {
|
|
vi.stubEnv("PAPERCLIP_RUNNER_E2E_PORT", "3100");
|
|
const api = new RunnerApi({} as APIRequestContext);
|
|
const fetcher = vi.fn(async (_input: RequestInfo | URL, _init?: RequestInit) => new Response("private failure body", { status: 403 })); vi.stubGlobal("fetch", fetcher);
|
|
await expect(api.getRunObservation("/api/heartbeat-runs/11111111-1111-4111-8111-111111111111/events?afterSeq=0&limit=1000")).rejects.toMatchObject({ name: "RunnerApiHttpError", status: 403 });
|
|
expect(fetcher).toHaveBeenCalledTimes(1);
|
|
expect(fetcher.mock.calls[0]![1]).toMatchObject({ headers: { "content-type": "application/json" } });
|
|
expect((fetcher.mock.calls[0]![1] as RequestInit).headers).not.toHaveProperty("Cookie");
|
|
});
|
|
it.each([
|
|
"https://foreign.invalid/api/heartbeat-runs/11111111-1111-4111-8111-111111111111",
|
|
"/api/companies/22222222-2222-4222-8222-222222222222/agents",
|
|
"/api/heartbeat-runs/11111111-1111-4111-8111-111111111111?token=anything",
|
|
"/api/heartbeat-runs/11111111-1111-4111-8111-111111111111/events?afterSeq=0&limit=10&extra=1",
|
|
"/api/companies/22222222-2222-4222-8222-222222222222/issues?limit=1000",
|
|
"/api/companies/22222222-2222-4222-8222-222222222222/issues?includeAgentCreatedTasks=true&limit=1000&q=x",
|
|
"/api/companies/22222222-2222-4222-8222-222222222222/activity?limit=1000&limit=1000",
|
|
])("rejects non-observation transport target %s before dispatch", async endpoint => {
|
|
vi.stubEnv("PAPERCLIP_RUNNER_E2E_PORT", "3100");
|
|
const fetcher = vi.fn(); vi.stubGlobal("fetch", fetcher);
|
|
await expect(new RunnerApi({} as APIRequestContext).getRunObservation(endpoint)).rejects.toThrow("allowlisted same-origin GET");
|
|
expect(fetcher).not.toHaveBeenCalled();
|
|
});
|
|
|
|
it.each([200, 204, 404, 409])("forwards the cleanup request timeout and preserves DELETE status %s", async status => {
|
|
vi.stubEnv("PAPERCLIP_RUNNER_E2E_PORT", "3100");
|
|
const remove = vi.fn(async () => response(status));
|
|
const api = new RunnerApi({ delete: remove } as unknown as APIRequestContext);
|
|
const result = api.delete("/api/environments/environment", { allowNotFound: true, timeout: 500 });
|
|
if (status === 409) await expect(result).rejects.toThrow("409");
|
|
else await expect(result).resolves.toBeUndefined();
|
|
expect(remove).toHaveBeenCalledExactlyOnceWith("/api/environments/environment", { timeout: 500 });
|
|
});
|