mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-13 07:12:00 +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>
133 lines
11 KiB
TypeScript
133 lines
11 KiB
TypeScript
import { CHAT_SETUP_HEALTH_ID, CHAT_SETUP_HEALTH_SUITE } from "./chat-setup-health.js";
|
|
import { afterEach, describe, expect, it, vi } from "vitest";
|
|
import { mkdtemp, readFile, rm } from "node:fs/promises";
|
|
import os from "node:os";
|
|
import path from "node:path";
|
|
import type { MatrixExecution } from "./types.js";
|
|
import { prepareChatRemoteHost, selectsRemoteChat } from "./chat-remote-host.js";
|
|
import { CHAT_OPENCODE_MODEL } from "./chat-runtime-profile.js";
|
|
import { chatCatalogScenarios } from "./chat-catalog-inventory.js";
|
|
const roots: string[] = [];
|
|
afterEach(async () => { await Promise.all(roots.splice(0).map(root => rm(root, { recursive: true, force: true }))); });
|
|
async function fixture(lifecycle = "spot") {
|
|
const root = await mkdtemp(path.join(os.tmpdir(), "remote-host-offline-")); roots.push(root);
|
|
const execution = { id: "chat-features.runner-opencode.local.create-backlog", suite: { id: "chat-features" }, profile: { id: "runner-opencode", provider: "opencode", model: CHAT_OPENCODE_MODEL, adapterType: "paperclip_runner", credential: "OPENROUTER_API_KEY" }, environment: { id: "local" }, task: { automaticRetryPolicy: "single_attempt" } } as unknown as MatrixExecution;
|
|
const env = { GITHUB_ACTIONS: "true", GITHUB_REPOSITORY: "paperclipai/paperclip", GITHUB_REF: "refs/heads/master", GITHUB_WORKFLOW_REF: "paperclipai/paperclip/.github/workflows/runner-full-stack-e2e.yml@refs/heads/master", GITHUB_RUN_ID: "123", GITHUB_RUN_ATTEMPT: "1", PAPERCLIP_RUNNER_E2E_SOURCE_SHA: "a".repeat(40), OPENROUTER_API_KEY: "synthetic-private-canary" };
|
|
const fetch = vi.fn(async (url: string | URL | Request) => {
|
|
const u = String(url);
|
|
if (u.endsWith("/api/token")) return new Response("synthetic-imds-token");
|
|
if (u.endsWith("/instance-life-cycle")) return new Response(lifecycle);
|
|
return Response.json({ instanceId: "i-1234567890abcdef0", imageId: "ami-1234567890abcdef0", region: "us-east-1", architecture: "x86_64", privateIp: "not-retained" });
|
|
});
|
|
return { executions: [execution], env, repositoryRoot: root, outputDirectory: root, dependencies: { fetch: fetch as typeof globalThis.fetch, platform: "linux", arch: "x64", sourceRevision: async () => "a".repeat(40) }, fetch,
|
|
receipt: async () => JSON.parse(await readFile(path.join(root, "chat-remote-host.json"), "utf8")) };
|
|
}
|
|
describe("remote-only chat host qualification without native/provider calls", () => {
|
|
it("admits the finite 244-case catalog only through its own trusted TestOps workflow", async () => {
|
|
const f = await fixture();
|
|
f.env.GITHUB_WORKFLOW_REF = "paperclipai/paperclip/.github/workflows/chat-catalog-testops.yml@refs/heads/master";
|
|
Object.assign(f.env, { PAPERCLIP_CHAT_COMPLETION_TESTOPS: "v1", ALLURE_JOB_RUN_ID: "77" });
|
|
f.executions = chatCatalogScenarios.map(row => {
|
|
const id = row.id === "tasks.create-backlog" ? "create-backlog" : row.id === "tasks.delegate-announcement" ? "delegate-announcement" : row.id;
|
|
return { ...f.executions[0]!, id: `chat-features.runner-opencode.local.${id}`, task: { ...f.executions[0]!.task, id } };
|
|
});
|
|
expect(f.executions).toHaveLength(244);
|
|
expect(await prepareChatRemoteHost(f)).toMatchObject({ maxAutomaticRetries: 0 });
|
|
const invalid = { ...f.executions[0]!, id: "chat-features.runner-opencode.local.foreign", task: { ...f.executions[0]!.task, id: "foreign" } };
|
|
f.fetch.mockClear(); f.executions = [invalid];
|
|
await expect(prepareChatRemoteHost(f)).rejects.toThrow("trusted-github-linux-source-required");
|
|
expect(f.fetch).not.toHaveBeenCalled();
|
|
});
|
|
it.each(["projects.project-work", "projects.project-defaults", "instance-operations.restart-service"])("admits only the bound TestOps pilot cell %s on the actual AWS host", async scenario => {
|
|
const f = await fixture();
|
|
f.env.GITHUB_WORKFLOW_REF = "paperclipai/paperclip/.github/workflows/chat-completion-testops.yml@refs/heads/master";
|
|
Object.assign(f.env, { PAPERCLIP_CHAT_COMPLETION_TESTOPS: "v1", ALLURE_JOB_RUN_ID: "77", GITHUB_RUN_ATTEMPT: "1" });
|
|
f.executions[0]!.id = `chat-features.runner-opencode.local.${scenario}`;
|
|
f.executions[0]!.task.id = scenario;
|
|
expect(await prepareChatRemoteHost(f)).toMatchObject({ databaseLane: "embedded-postgres", maxAutomaticRetries: 0 });
|
|
expect(await f.receipt()).toMatchObject({ workflowRef: f.env.GITHUB_WORKFLOW_REF, evaluationAdmitted: false,
|
|
testOps: { admission: "v1", allureJobRunId: "77" } });
|
|
});
|
|
it.each(["foreign-cell", "wrong-task-id", "missing-admission", "missing-jobrun", "replay", "candidate-workflow"])("rejects TestOps %s before AWS metadata or inference", async kind => {
|
|
const f = await fixture();
|
|
f.env.GITHUB_WORKFLOW_REF = "paperclipai/paperclip/.github/workflows/chat-completion-testops.yml@refs/heads/master";
|
|
const extra = { PAPERCLIP_CHAT_COMPLETION_TESTOPS: "v1", ALLURE_JOB_RUN_ID: "77" };
|
|
Object.assign(f.env, extra); f.executions[0]!.id = "chat-features.runner-opencode.local.projects.project-work";
|
|
f.executions[0]!.task.id = "projects.project-work";
|
|
if (kind === "foreign-cell") f.executions[0]!.id = "chat-features.runner-opencode.local.tasks.delete";
|
|
if (kind === "wrong-task-id") f.executions[0]!.task.id = "foreign";
|
|
if (kind === "missing-admission") Object.assign(f.env, { PAPERCLIP_CHAT_COMPLETION_TESTOPS: undefined });
|
|
if (kind === "missing-jobrun") Object.assign(f.env, { ALLURE_JOB_RUN_ID: undefined });
|
|
if (kind === "replay") f.env.GITHUB_RUN_ATTEMPT = "2";
|
|
if (kind === "candidate-workflow") f.env.GITHUB_WORKFLOW_REF = f.env.GITHUB_WORKFLOW_REF.replace("refs/heads/master", "refs/heads/candidate");
|
|
await expect(prepareChatRemoteHost(f)).rejects.toThrow("trusted-github-linux-source-required");
|
|
expect(f.fetch).not.toHaveBeenCalled();
|
|
});
|
|
it.each(["00077", "9007199254740992", "12345678901234567"])("rejects the same unsafe job identity %s as the dedicated dispatcher", async jobRunId => {
|
|
const f = await fixture(); f.env.GITHUB_WORKFLOW_REF = "paperclipai/paperclip/.github/workflows/chat-completion-testops.yml@refs/heads/master";
|
|
Object.assign(f.env, { PAPERCLIP_CHAT_COMPLETION_TESTOPS: "v1", ALLURE_JOB_RUN_ID: jobRunId });
|
|
f.executions[0]!.task.id = "projects.project-work"; f.executions[0]!.id = `chat-features.runner-opencode.local.${f.executions[0]!.task.id}`;
|
|
await expect(prepareChatRemoteHost(f)).rejects.toThrow("trusted-github-linux-source-required"); expect(f.fetch).not.toHaveBeenCalled();
|
|
});
|
|
it("admits the maximum safe job identity and retains it exactly", async () => {
|
|
const f = await fixture(); f.env.GITHUB_WORKFLOW_REF = "paperclipai/paperclip/.github/workflows/chat-completion-testops.yml@refs/heads/master";
|
|
Object.assign(f.env, { PAPERCLIP_CHAT_COMPLETION_TESTOPS: "v1", ALLURE_JOB_RUN_ID: "9007199254740991" });
|
|
f.executions[0]!.task.id = "projects.project-work"; f.executions[0]!.id = `chat-features.runner-opencode.local.${f.executions[0]!.task.id}`;
|
|
await prepareChatRemoteHost(f); expect(await f.receipt()).toMatchObject({ testOps: { admission: "v1", allureJobRunId: "9007199254740991" } });
|
|
});
|
|
it("requires real Spot metadata and retains only closed source-bound host proof", async () => {
|
|
const f = await fixture(); const policy = await prepareChatRemoteHost(f); const receipt = await f.receipt();
|
|
expect(policy).toMatchObject({ databaseLane: "embedded-postgres", maxAutomaticRetries: 0, maxParallel: 2 });
|
|
expect(receipt).toMatchObject({ status: "qualified-host-only", spotVerified: true, evaluationAdmitted: false, allocationSpotOnlyClaimed: false, sourceRevision: "a".repeat(40) });
|
|
expect(JSON.stringify(receipt)).not.toMatch(/synthetic-private-canary|synthetic-imds-token|privateIp|not-retained/);
|
|
expect(f.fetch.mock.calls.every(([url]) => String(url).startsWith("http://169.254.169.254/latest/"))).toBe(true);
|
|
expect(receipt.databaseLane).toBe("embedded-postgres");
|
|
expect(receipt).not.toHaveProperty("postgres");
|
|
expect(receipt).not.toHaveProperty("dockerEngineSha256");
|
|
});
|
|
it("accepts the user-authorized fleet fallback while truthfully recording On-Demand", async () => {
|
|
const f = await fixture("on-demand"); await prepareChatRemoteHost(f);
|
|
expect(await f.receipt()).toMatchObject({ lifecycle: "on-demand", status: "qualified-host-only", spotVerified: false, hostPolicy: "aws-fleet-spot-or-on-demand", evaluationAdmitted: false });
|
|
});
|
|
it.each(["", "unknown", "preemptible"])("refuses unproven host lifecycle %s", async lifecycle => {
|
|
const f = await fixture(lifecycle); await expect(prepareChatRemoteHost(f)).rejects.toThrow("unknown-host-lifecycle-refused");
|
|
expect(await f.receipt()).toMatchObject({ lifecycle, status: "blocked", evaluationAdmitted: false });
|
|
});
|
|
it.each(["local", "branch-workflow", "wrong-source", "wrong-model", "automatic-retry", "foreign-repository"])("fails closed on %s before metadata", async kind => {
|
|
const f = await fixture();
|
|
if (kind === "local") f.dependencies.platform = "darwin";
|
|
if (kind === "branch-workflow") f.env.GITHUB_WORKFLOW_REF = "untrusted@refs/heads/branch";
|
|
if (kind === "wrong-source") f.env.PAPERCLIP_RUNNER_E2E_SOURCE_SHA = "b".repeat(40);
|
|
if (kind === "wrong-model") f.executions[0]!.profile.model = "other";
|
|
if (kind === "automatic-retry") f.executions[0]!.task.automaticRetryPolicy = undefined;
|
|
if (kind === "foreign-repository") f.env.GITHUB_REPOSITORY = "foreign/repository";
|
|
await expect(prepareChatRemoteHost(f)).rejects.toThrow("trusted-github-linux-source-required"); expect(f.fetch).not.toHaveBeenCalled();
|
|
});
|
|
it("records missing actual AWS identity without admitting an evaluation", async () => {
|
|
const f = await fixture(); f.dependencies.fetch = vi.fn(async () => new Response("", { status: 403 }));
|
|
await expect(prepareChatRemoteHost(f)).rejects.toThrow("aws-instance-identity-unavailable");
|
|
expect(await f.receipt()).toMatchObject({ status: "blocked", evaluationAdmitted: false, databaseLane: "embedded-postgres" });
|
|
});
|
|
it.each(["architecture", "region", "instanceId"])("rejects wrong AWS %s", async field => {
|
|
const f = await fixture(), original = f.dependencies.fetch;
|
|
f.dependencies.fetch = vi.fn(async (url, init) => {
|
|
const response = await original(url, init);
|
|
if (!String(url).endsWith("/document")) return response;
|
|
const identity = await response.json(); identity[field] = "wrong"; return Response.json(identity);
|
|
});
|
|
await expect(prepareChatRemoteHost(f)).rejects.toThrow("aws-instance-identity-unavailable");
|
|
});
|
|
it("keeps the one exact health cell on the same AWS/ordinary embedded database policy", async () => {
|
|
const f = await fixture("on-demand"); f.executions[0]!.id = CHAT_SETUP_HEALTH_ID; f.executions[0]!.suite.id = CHAT_SETUP_HEALTH_SUITE;
|
|
expect(selectsRemoteChat(f.executions)).toBe(true);
|
|
expect(await prepareChatRemoteHost(f)).toMatchObject({ databaseLane: "embedded-postgres", maxAutomaticRetries: 0, maxParallel: 2 });
|
|
expect(await f.receipt()).not.toHaveProperty("dockerEngineSha256");
|
|
f.executions[0]!.id = "chat-setup-health.runner-opencode.local.wrong";
|
|
expect(selectsRemoteChat(f.executions)).toBe(false);
|
|
});
|
|
it("does not change historical/other profiles or require IMDS for their discovery", async () => {
|
|
const f = await fixture(); f.executions[0]!.profile.id = "runner-codex-subscription";
|
|
expect(selectsRemoteChat(f.executions)).toBe(false); expect(await prepareChatRemoteHost(f)).toBeUndefined(); expect(f.fetch).not.toHaveBeenCalled();
|
|
});
|
|
});
|