mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-14 00:26:28 +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>
44 lines
5.7 KiB
TypeScript
44 lines
5.7 KiB
TypeScript
import type { ChatCatalogCase, ChatCatalogContext, ChatCatalogSetupBoundary, NativeStopBoundaryReceipt } from "./chat-catalog-contract.js";
|
|
export function validNativeStopDeclaration(value: Omit<Pick<ChatCatalogCase, "scenarioId" | "canonicalBoundary" | "authentication" | "journeyKind" | "expectedHumanMessageCount" | "setupHumanMessageCount" | "queryOutcome" | "queryBindings" | "queryAttachments">, "journeyKind"> & { journeyKind?: string }) {
|
|
return value.canonicalBoundary === "native-stop-cleanup" && value.scenarioId === "recovery.post-stop-message" && value.authentication === "human-session"
|
|
&& (value.journeyKind ?? "agent-chat") === "agent-chat" && value.expectedHumanMessageCount === 2 && value.setupHumanMessageCount === 1
|
|
&& (value.queryOutcome ?? "persisted") === "persisted" && !value.queryBindings?.length && !value.queryAttachments?.length;
|
|
}
|
|
export function assertNativeStopCallback(implementation: ChatCatalogCase) {
|
|
if (implementation.canonicalBoundary ? !validNativeStopDeclaration(implementation) || typeof implementation.beforeCanonicalSend !== "function" : implementation.beforeCanonicalSend !== undefined)
|
|
throw new Error("Missing, wrong-case or undeclared native Stop canonical-boundary callback");
|
|
}
|
|
// Production native-session-executor writes this intent in its cancellation transaction,
|
|
// then only writes dispatchState=acknowledged/acknowledgedAt after actual cleanup.
|
|
// A request receipt or a cancelled status alone does not establish this interval.
|
|
export function nativeStopPendingReceiptMatches(input: { receipt: NativeStopBoundaryReceipt; run: { id?: string; companyId?: string; nativeIssueId?: string; agentId?: string; runtimeMode?: string; status?: string; resultJson?: { nativeCancellation?: Record<string, unknown> } }; companyId: string; sourceIssueId: string; assignedAgentId: string; setupRunIds: string[] }) {
|
|
const { receipt: r, run } = input, intent = run.resultJson?.nativeCancellation;
|
|
return !!r && Object.keys(r).every(key => ["runId", "issueId", "intentId", "intentRecordedAt"].includes(key)) && [r.runId, r.issueId].every(id => typeof id === "string" && /^[a-f0-9-]{36}$/i.test(id))
|
|
// Production intent ids are `native-cancellation:<uuid>` (native-cancellation-request.ts).
|
|
&& typeof r.intentId === "string" && /^native-cancellation:[a-f0-9-]{36}$/i.test(r.intentId)
|
|
&& Number.isFinite(Date.parse(r.intentRecordedAt)) && input.setupRunIds.includes(r.runId) && run.id === r.runId && run.companyId === input.companyId && run.nativeIssueId === r.issueId
|
|
&& r.issueId === input.sourceIssueId && run.agentId === input.assignedAgentId && run.runtimeMode === "native" && run.status === "running" && intent?.schema === "paperclip.native-cancellation.v1" && intent.companyId === input.companyId && intent.runId === r.runId && intent.issueId === r.issueId
|
|
&& intent.intentId === r.intentId && intent.recordedAt === r.intentRecordedAt && intent.scope === "run" && intent.dispatchState === "pending" && intent.dispatched === false && intent.acknowledgedAt == null;
|
|
}
|
|
export async function prepareNativeStopCanonicalBoundary(ctx: ChatCatalogContext, implementation: ChatCatalogCase, state: Record<string, unknown>, canonicalQuery: string, setup: ChatCatalogSetupBoundary) {
|
|
assertNativeStopCallback(implementation); if (!implementation.canonicalBoundary) return;
|
|
if (canonicalQuery !== ctx.scenario.query || setup.humanCommentIds.length !== 1 || setup.runIds.length !== 1) throw new Error("Native Stop boundary cannot alter the canonical query or introduce setup chat turns");
|
|
const source = ctx.issue(); if (!source) throw new Error("Native Stop boundary requires the existing actual source chat");
|
|
const sourceURL = ctx.input.page.url(), originalQuery = ctx.scenario.query;
|
|
const before = await ctx.input.api.get<Array<{ id: string; body: string; authorAgentId?: string }>>(`/api/issues/${source.id}/comments?order=asc`);
|
|
const retainedHuman = before.filter(row => !row.authorAgentId);
|
|
if (retainedHuman.length !== 1 || retainedHuman[0]?.id !== setup.humanCommentIds[0]) throw new Error("Native Stop boundary requires exactly the declared genuine setup human message");
|
|
const receipt = await implementation.beforeCanonicalSend!(ctx, state);
|
|
const [run, comments, runs, actualSource] = await Promise.all([
|
|
ctx.input.api.get<Parameters<typeof nativeStopPendingReceiptMatches>[0]["run"]>(`/api/heartbeat-runs/${receipt.runId}`),
|
|
ctx.input.api.get<Array<{ id: string; body: string; authorAgentId?: string }>>(`/api/issues/${source.id}/comments?order=asc`), ctx.allRuns(),
|
|
ctx.input.api.get<{ id: string; companyId: string; assigneeAgentId?: string | null; conversationAgentId?: string | null }>(`/api/issues/${source.id}`),
|
|
]);
|
|
const assignedAgentId = actualSource.conversationAgentId ?? actualSource.assigneeAgentId;
|
|
if (actualSource.id !== source.id || actualSource.companyId !== ctx.input.fixtures.company.id || !assignedAgentId || !nativeStopPendingReceiptMatches({ receipt, run, companyId: ctx.input.fixtures.company.id, sourceIssueId: source.id, assignedAgentId, setupRunIds: setup.runIds }) || ctx.issue()?.id !== source.id || ctx.input.page.url() !== sourceURL
|
|
|| ctx.scenario.query !== originalQuery || canonicalQuery !== originalQuery || JSON.stringify(comments.filter(row => !row.authorAgentId)) !== JSON.stringify(retainedHuman)
|
|
|| runs.length !== setup.runIds.length || runs.some(row => !setup.runIds.includes(row.id))) throw new Error("Native Stop boundary changed source/query/history or missed the genuine pending cleanup interval");
|
|
state.nativeStopBoundary = receipt;
|
|
await ctx.input.evidence("chat-catalog-native-stop-boundary.json", { kind: implementation.canonicalBoundary, sourceIssueId: source.id, ...receipt, canonicalQueryUnchanged: true, retainedSetupHumanCommentId: retainedHuman[0]!.id, extraHumanMessages: 0, extraRuns: 0 });
|
|
}
|