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>
22 lines
6.6 KiB
TypeScript
22 lines
6.6 KiB
TypeScript
import { expect, it } from "vitest";
|
|
import { actualOriginalEmailSend, actualUncertainEmailReads, gradeUnknownEmail, uncertainEmailReply } from "./runtime-email.js";
|
|
function valid() { return { companyId: "company", leadId: "ceo", scenarioId: "recovery.unknown-email-outcome", origin: "http://127.0.0.1:42123", reply: "", agents: [], tasks: [{ id: "email-task" }, { id: "mail-work" }], activity: [], links: [], humanQueries: [], runs: [1, 2, 3, 4].map(id => ({ id: String(id), agentId: "ceo", status: "succeeded" })), state: { actualNormalTaskSendBound: true, actualOriginalUncertain: true, setupSendBound: true, faultOriginalBinding: true, outageActiveDuringCanonical: true, originalUncertaintyRecorded: true, canonicalReadBound: true, canonicalReply: "The original provider receipt is unavailable and the send remains uncertain. I will check the existing publication rather than resend a possibly delivered email.", noNewAgentPublication: true, directMatchedSentCount: 1, originalProviderReceiptMatched: true, outageReleased: true, reconciliationBound: true, resolutionActorBound: true, finalOutcome: "sent", confirmationBound: true, confirmationReply: "The original email was accepted and its saved publication is reconciled with the provider message receipt. There is no proof of recipient delivery, and no second send was made.", finalDirectMatchedSentCount: 1, sameOriginalRequestRetained: true, sameOriginalConversationRetained: true } }; }
|
|
it("accepts real original uncertainty, independent one-send proof and authorized same-publication reconciliation", () => expect(gradeUnknownEmail(valid()).passed).toBe(true));
|
|
it("accepts natural uncertainty and accepted-send paraphrases", () => { const o = valid(); o.state.canonicalReply = "It may already have reached the provider. I cannot confirm the outcome, so I checked the original send record and will avoid a duplicate while we reconcile the receipt."; o.state.confirmationReply = "The provider's original message is now verified as sent; the same conversation remains linked. Recipient delivery is still unconfirmed."; expect(gradeUnknownEmail(o).passed).toBe(true); });
|
|
it("rejects another actual email, a new agent publication, or loss of original uncertainty", () => { const o = valid(); o.state.directMatchedSentCount = 2; expect(gradeUnknownEmail(o).passed).toBe(false); const p = valid(); p.state.noNewAgentPublication = false; expect(gradeUnknownEmail(p).passed).toBe(false); const q = valid(); q.state.originalUncertaintyRecorded = false; expect(gradeUnknownEmail(q).passed).toBe(false); });
|
|
it("rejects automatic resend or certainty unsupported by the actual uncertain record", () => { expect(uncertainEmailReply("It is uncertain, but I will resend it now while we check the provider receipt.")).toBe(false); expect(uncertainEmailReply("The provider receipt is unavailable, but it definitely failed. I will resend it.")).toBe(false); });
|
|
it("rejects claims of delivery from accepted-send evidence and wrong final reply phase", () => { const o = valid(); o.state.confirmationReply += " It was delivered."; expect(gradeUnknownEmail(o).passed).toBe(false); const p = valid(); p.state.confirmationBound = false; expect(gradeUnknownEmail(p).passed).toBe(false); });
|
|
const expected = { companyId: "company", endpointId: "endpoint", publicationId: "publication", issueId: "email-task" };
|
|
function actualRead(api = false) { const name = api ? "call_api" : "agentmail_delivery", input = api ? { operationId: "GET /api/companies/{companyId}/email/deliveries/{publicationId}", pathParams: { companyId: "company", publicationId: "publication" } } : { publicationId: "publication" }, receipt = { id: "publication", issueId: "email-task", outcome: "uncertain", providerMessageId: null, request: { endpointId: "endpoint" } }; return [{ seq: 1, eventType: "item.started", payload: { prpEvent: { payload: { item: { id: "read", type: "tool_use", name, input } } } } }, { seq: 2, eventType: "item.completed", payload: { prpEvent: { payload: { item: { id: "read", type: "tool_result", result: api ? { ok: true, status: 200, data: receipt } : receipt } } } } }, { seq: 3, eventType: "tool.execution.completed", payload: { prpEvent: { payload: { executionId: "read", name, status: "completed" } } } }]; }
|
|
it("accepts actual completed connector and equivalent scoped API receipt reads", () => { expect(actualUncertainEmailReads(actualRead(), expected)).toHaveLength(1); expect(actualUncertainEmailReads(actualRead(true), expected)).toHaveLength(1); });
|
|
it("rejects wrong receipt identity, uncompleted or wrong-tool results, and a write before inspection", () => { expect(actualUncertainEmailReads(actualRead(), { ...expected, publicationId: "other" })).toHaveLength(0); expect(actualUncertainEmailReads(actualRead().slice(0, 2), expected)).toHaveLength(0); const wrong = actualRead(); wrong[2]!.payload.prpEvent.payload = { executionId: "read", name: "different_tool", status: "completed" }; expect(actualUncertainEmailReads(wrong, expected)).toHaveLength(0); const early = { seq: 0, eventType: "item.started", payload: { prpEvent: { payload: { item: { id: "send", type: "tool_use", name: "agentmail_send", input: {} } } } } }; expect(actualUncertainEmailReads([early, ...actualRead()], expected)).toHaveLength(0); });
|
|
|
|
it("accepts uncertainty about delivery while rejecting appended affirmative delivery", () => { expect(uncertainEmailReply("I cannot confirm whether it was delivered; I checked the original provider receipt and will avoid another send.")).toBe(true); expect(uncertainEmailReply("The provider outcome is uncertain. It was delivered.")).toBe(false); });
|
|
|
|
it("requires genuine normal-task send attribution rather than chat execution or fixture-only state", () => {
|
|
const target = { companyId: "company", taskId: "mail-work", endpointId: "endpoint", publicationId: "publication" };
|
|
const events = [{ eventType: "item.started", payload: { prpEvent: { payload: { item: { id: "send", type: "tool_use", name: "agentmail_send", input: { request: { parentIssueId: "mail-work", endpointId: "endpoint", idempotencyKey: "publication" } } } } } } }, { eventType: "item.completed", payload: { prpEvent: { payload: { item: { id: "send", type: "tool_result", result: { id: "publication", issueId: "email-child" } } } } } }, { eventType: "tool.execution.completed", payload: { prpEvent: { payload: { executionId: "send", name: "agentmail_send", status: "completed" } } } }];
|
|
expect(actualOriginalEmailSend(events, target)).toBe(true); expect(actualOriginalEmailSend(events, { ...target, taskId: "agent-chat" })).toBe(false); expect(actualOriginalEmailSend(events.slice(0, 2), target)).toBe(false);
|
|
const o = valid(); o.state.actualNormalTaskSendBound = false; expect(gradeUnknownEmail(o).passed).toBe(false);
|
|
});
|