Files
PaperClipAI/tests/runner-e2e/chat-native-stop-boundary.test.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

56 lines
8.4 KiB
TypeScript

import { expect, it, vi } from "vitest";
import type { ChatCatalogCase, ChatCatalogContext, NativeStopBoundaryReceipt } from "./chat-catalog-contract.js";
import { loadChatCatalogImplementation, type ChatCatalogDescriptor } from "./chat-catalog-registry.js";
import { assertNativeStopCallback, nativeStopPendingReceiptMatches, prepareNativeStopCanonicalBoundary, validNativeStopDeclaration } from "./chat-native-stop-boundary.js";
const receipt: NativeStopBoundaryReceipt = { runId: "11111111-1111-4111-8111-111111111111", issueId: "22222222-2222-4222-8222-222222222222", intentId: "native-cancellation:33333333-3333-4333-8333-333333333333", intentRecordedAt: "2026-10-26T12:00:00.000Z" };
const metadata: ChatCatalogDescriptor = { scenarioId: "recovery.post-stop-message", canonicalBoundary: "native-stop-cleanup", authentication: "human-session", sourceFiles: ["runtime-post-stop.ts"], modulePath: "./chat-catalog-domains/runtime-post-stop.js", exportName: "cases", oracleVersion: "control", expectedRunCount: 3, expectedHumanMessageCount: 2, setupHumanMessageCount: 1, turnTimeoutMs: 120000, attemptTimeoutMs: 600000 };
function fixture() {
const run = { id: receipt.runId, companyId: "company", nativeIssueId: receipt.issueId, agentId: "lead", runtimeMode: "native", status: "running", resultJson: { nativeCancellation: { schema: "paperclip.native-cancellation.v1", companyId: "company", runId: receipt.runId, issueId: receipt.issueId, scope: "run", intentId: receipt.intentId, recordedAt: receipt.intentRecordedAt, dispatchState: "pending", dispatched: false } } };
const comments: Array<{ id: string; body: string; authorAgentId?: string }> = [{ id: "setup-human", body: "Build the original website after the bounded wait." }], runs = [{ id: receipt.runId }]; let url = "http://fixture/source";
const ctx = { scenario: { id: metadata.scenarioId, query: "Make the page simpler." }, issue: () => ({ id: receipt.issueId, assigneeAgentId: "lead" }), allRuns: async () => runs,
input: { fixtures: { company: { id: "company" } }, page: { url: () => url }, evidence: vi.fn(async () => {}), api: { get: async (path: string) => path.startsWith("/api/heartbeat-runs/") ? run : path.endsWith("/comments?order=asc") ? structuredClone(comments) : { id: receipt.issueId, companyId: "company", conversationAgentId: "lead" } } } } as unknown as ChatCatalogContext;
const implementation = { ...metadata, setup: async () => ({}), observe: async () => ({}), grade: () => ({ passed: false, checks: [] }), verifyResult: async () => {}, beforeCanonicalSend: vi.fn(async () => ({ ...receipt })) } as unknown as ChatCatalogCase;
return { run, comments, runs, ctx, implementation, move: () => { url += "/other"; }, setup: { humanCommentIds: ["setup-human"], runIds: [receipt.runId] } };
}
it("admits only the exact unchanged human post-stop boundary declaration", () => {
expect(validNativeStopDeclaration(metadata)).toBe(true);
for (const patch of [{ scenarioId: "recovery.explain-stall" }, { journeyKind: "browser-boundary" }, { authentication: "local-trusted" }, { expectedHumanMessageCount: 1 }, { setupHumanMessageCount: 0 }, { queryOutcome: "rejected" }, { queryAttachments: [{ stateKey: "file", filename: "brief.md" }] }]) expect(validNativeStopDeclaration({ ...metadata, ...patch } as typeof metadata)).toBe(false);
expect(() => assertNativeStopCallback({ ...fixture().implementation, beforeCanonicalSend: undefined })).toThrow("callback");
expect(() => assertNativeStopCallback({ ...fixture().implementation, canonicalBoundary: undefined })).toThrow("callback");
});
it("preflight rejects declaration drift and absent or hidden callbacks before any admission", async () => {
const f = fixture(), load = (row: ChatCatalogCase, descriptor = metadata) => loadChatCatalogImplementation(descriptor, async () => ({ cases: [row] }));
await expect(load(f.implementation)).resolves.toBe(f.implementation);
await expect(load({ ...f.implementation, beforeCanonicalSend: undefined })).rejects.toThrow("callback");
await expect(load({ ...f.implementation, canonicalBoundary: undefined })).rejects.toThrow("drift");
await expect(load({ ...f.implementation, canonicalBoundary: undefined }, { ...metadata, canonicalBoundary: undefined })).rejects.toThrow("callback");
});
it("requires actual same-company admitted running setup work and a pending unacknowledged persisted intent", () => {
const input = () => { const f = fixture(); return { receipt: { ...receipt }, run: f.run, companyId: "company", sourceIssueId: receipt.issueId, assignedAgentId: "lead", setupRunIds: [receipt.runId] }; };
expect(nativeStopPendingReceiptMatches(input())).toBe(true);
for (const change of [(v: ReturnType<typeof input>) => { v.run.companyId = "foreign"; }, (v: ReturnType<typeof input>) => { v.run.status = "cancelled"; }, (v: ReturnType<typeof input>) => { v.sourceIssueId = "another-source"; }, (v: ReturnType<typeof input>) => { v.run.agentId = "another-agent"; }, (v: ReturnType<typeof input>) => { v.run.runtimeMode = "legacy"; }, (v: ReturnType<typeof input>) => { v.setupRunIds = []; }, (v: ReturnType<typeof input>) => { v.run.resultJson.nativeCancellation.dispatchState = "acknowledged"; }, (v: ReturnType<typeof input>) => { v.run.resultJson.nativeCancellation.dispatched = true; }, (v: ReturnType<typeof input>) => { v.receipt.intentId = receipt.runId; }, (v: ReturnType<typeof input>) => { v.receipt.intentRecordedAt = "invalid"; }]) { const value = input(); change(value); expect(nativeStopPendingReceiptMatches(value)).toBe(false); }
});
it("keeps the canonical human message untouched and rejects hidden turns, source navigation or extra admissions", async () => {
const success = fixture(), state = {};
await prepareNativeStopCanonicalBoundary(success.ctx, success.implementation, state, success.ctx.scenario.query, success.setup);
expect(state).toEqual({ nativeStopBoundary: receipt }); expect(success.implementation.beforeCanonicalSend).toHaveBeenCalledOnce();
for (const change of [(f: ReturnType<typeof fixture>) => { f.comments.push({ id: "hidden-human", body: "Hidden direction" }); }, (f: ReturnType<typeof fixture>) => { f.runs.push({ id: "extra-run" }); }, (f: ReturnType<typeof fixture>) => { f.comments[0]!.body = "Altered setup direction"; }, (f: ReturnType<typeof fixture>) => { f.move(); }, (f: ReturnType<typeof fixture>) => { f.ctx.scenario.query += " Hurry"; }]) {
const f = fixture(); f.implementation.beforeCanonicalSend = async () => { change(f); return receipt; };
await expect(prepareNativeStopCanonicalBoundary(f.ctx, f.implementation, {}, f.ctx.scenario.query, f.setup)).rejects.toThrow("changed source/query/history");
}
const f = fixture(); await expect(prepareNativeStopCanonicalBoundary(f.ctx, f.implementation, {}, "Alternate direction", f.setup)).rejects.toThrow("alter"); expect(f.implementation.beforeCanonicalSend).not.toHaveBeenCalled();
});
it("does no boundary work in an ordinary undeclared catalog scenario", async () => {
const f = fixture(); await prepareNativeStopCanonicalBoundary(f.ctx, { ...f.implementation, canonicalBoundary: undefined, beforeCanonicalSend: undefined }, {}, "Exact ordinary query", f.setup);
expect(f.ctx.input.evidence).not.toHaveBeenCalled();
});
it("preflights the finite Slack distractor callback only for its declared reply thread", async () => {
const descriptor: ChatCatalogDescriptor = { ...metadata, scenarioId: "chat-channels.reply-in-thread", canonicalBoundary: undefined, journeyKind: "native-slack-thread", slackIngress: { kind: "native-slack-thread", maximumProviderCalls: 20, ingressTimeoutMs: 120000, setupThread: "same-agent-distractor" }, expectedRunCount: 4, minimumRunCount: 2, expectedHumanMessageCount: 1, setupHumanMessageCount: 0 };
const row = { ...fixture().implementation, ...descriptor, beforeCanonicalSend: undefined, dispatchSlackQuery: async () => ({}), dispatchSlackOtherThread: async () => ({}) } as unknown as ChatCatalogCase;
const load = (implementation: ChatCatalogCase, selected = descriptor) => loadChatCatalogImplementation(selected, async () => ({ cases: [implementation] }));
await expect(load(row)).resolves.toBe(row);
await expect(load({ ...row, dispatchSlackOtherThread: undefined })).rejects.toThrow("other-thread callback");
const ordinary = { ...metadata, canonicalBoundary: undefined };
await expect(load({ ...fixture().implementation, canonicalBoundary: undefined, beforeCanonicalSend: undefined, dispatchSlackOtherThread: row.dispatchSlackOtherThread }, ordinary)).rejects.toThrow("other-thread callback");
});