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>
44 lines
6.3 KiB
TypeScript
44 lines
6.3 KiB
TypeScript
import { randomUUID } from "node:crypto";
|
|
import { eq } from "drizzle-orm";
|
|
import { afterAll, beforeAll, expect, it } from "vitest";
|
|
import { agents, companies, completionContracts, createDb, heartbeatRuns, issues } from "@paperclipai/db";
|
|
import { startRecoveryPostgres } from "../../__tests__/helpers/recovery-postgres.js";
|
|
import { nativeCompletionFeedback } from "./native-completion-feedback.js";
|
|
import type { PrpStructuredRunResult } from "../../vendor/paperclip-runner/index.js";
|
|
|
|
let temporary: Awaited<ReturnType<typeof startRecoveryPostgres>>;
|
|
let db: ReturnType<typeof createDb>;
|
|
beforeAll(async () => { temporary = await startRecoveryPostgres("failed-action-chat-feedback-"); db = createDb(temporary.connectionString); });
|
|
afterAll(async () => { await temporary?.cleanup(); });
|
|
it("compares truthful failed-action chat reports with ordinary-task and false-success controls", async () => {
|
|
const companyId = randomUUID(), agentId = randomUUID();
|
|
await db.insert(companies).values({ id: companyId, name: "Feedback control", issuePrefix: `C${companyId.slice(0, 7)}` });
|
|
await db.insert(agents).values({ id: agentId, companyId, name: "CEO", adapterType: "paperclip_runner", status: "active" });
|
|
async function fixture(conversation: boolean) {
|
|
const issueId = randomUUID(), runId = randomUUID(), contractId = randomUUID();
|
|
await db.insert(issues).values({ id: issueId, companyId, title: conversation ? "Chat with CEO" : "Update launch page", status: "in_progress", assigneeAgentId: agentId,
|
|
...(conversation ? { conversationAgentId: agentId, conversationUserId: "control-user", conversationState: "active" } : {}) });
|
|
await db.insert(heartbeatRuns).values({ id: runId, companyId, agentId, nativeIssueId: issueId, status: "running", runtimeMode: "native", completionContractId: contractId,
|
|
contextSnapshot: { issueId, conversationMode: conversation, source: "tool_action_review", toolAction: { invocationId: randomUUID(), actionRequestId: randomUUID(), executionStatus: "failed", decision: "accepted", error: "Synthetic write failed; page unchanged", instructions: "The approved action did not complete successfully. Do not automatically replay it; inspect the recorded outcome first." } } });
|
|
await db.insert(completionContracts).values({ id: contractId, companyId, issueId, revision: 4, schemaVersion: "paperclip.completion-contract.v1", policyVersion: "phase6-v1", risk: "standard", completionAuthority: "server_arbiter", incompleteCriteriaPolicy: "preserve_non_terminal", canonicalSha256: randomUUID(), createdByActorType: "system", createdByActorId: "control", contractJson: { revision: "4", criteria: [{ id: "objective", requirement: conversation ? "Address this request in the current authorized conversation response: Respond to the current conversation context." : "Update launch page successfully." }] } });
|
|
await db.update(issues).set({ executionRunId: runId }).where(eq(issues.id, issueId));
|
|
return runId;
|
|
}
|
|
const report: PrpStructuredRunResult = { schema: "paperclip.run_result.v1", reportedWorkDisposition: "done", summary: "The approved write failed: Synthetic write failed; page unchanged. The page still reads Controlled launch page. No retry executed. A retry requires a new user request.", completionClaim: { contractRevision: "4", objectiveSatisfied: true, criteria: [{ criterionId: "objective", status: "satisfied", evidenceRefs: [] }], remainingWork: [{ description: "The external page update remains incomplete; retry only on user request.", blocksCompletion: false }] }, evidence: [], verification: [{ commandOrCheck: "Approved external write", status: "failed", detail: "Synthetic write failed; page unchanged" }], attentionRequests: [], artifacts: [] };
|
|
const chat = await fixture(true), ordinary = await fixture(false);
|
|
await expect(nativeCompletionFeedback(db, chat, report)).rejects.toThrow(/Rejection fields: verification\.failed\.[\s\S]*Do not automatically replay the action or request approval again/);
|
|
for (const patch of [{ source: "issue.comment" }, { toolAction: { executionStatus: "succeeded", invocationId: randomUUID(), actionRequestId: randomUUID() } }, { toolAction: { executionStatus: "failed" } }]) {
|
|
const [row] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.id, chat));
|
|
const original = row!.contextSnapshot;
|
|
await db.update(heartbeatRuns).set({ contextSnapshot: { ...original, ...patch } }).where(eq(heartbeatRuns.id, chat));
|
|
await expect(nativeCompletionFeedback(db, chat, report)).rejects.toThrow(/^The done report includes unfinished work, failed verification, or an outstanding decision\. Finish the work or report the concrete blocker\/reviewer request\. No human completion approval was created\.$/);
|
|
await db.update(heartbeatRuns).set({ contextSnapshot: original }).where(eq(heartbeatRuns.id, chat));
|
|
}
|
|
await expect(nativeCompletionFeedback(db, ordinary, report)).rejects.toThrow(/^The done report includes unfinished work, failed verification, or an outstanding decision\. Finish the work or report the concrete blocker\/reviewer request\. No human completion approval was created\.$/);
|
|
await expect(nativeCompletionFeedback(db, chat, { ...report, summary: "The page update succeeded." })).rejects.toThrow("failed verification");
|
|
await expect(nativeCompletionFeedback(db, chat, { ...report, verification: [], completionClaim: { ...report.completionClaim, remainingWork: [{ description: "Retry silently before completion", blocksCompletion: true }] } })).rejects.toThrow("Rejection fields: completionClaim.remainingWork.blocksCompletion.");
|
|
await expect(nativeCompletionFeedback(db, ordinary, { ...report, verification: [], completionClaim: { ...report.completionClaim, objectiveSatisfied: false, criteria: [{ criterionId: "objective", status: "not_satisfied", evidenceRefs: [] }] } })).rejects.toThrow("unfinished work");
|
|
await expect(nativeCompletionFeedback(db, chat, { ...report, verification: [{ commandOrCheck: "External write completion", status: "not_run", reasonCode: "external_service_unavailable", detail: "Attempted approved write returned Synthetic write failed; page unchanged." }] })).resolves.toContain("accepted");
|
|
await expect(nativeCompletionFeedback(db, chat, { ...report, verification: [{ commandOrCheck: "Read page after failed attempt", status: "passed", detail: "Verified Controlled launch page is unchanged; external write failed and no retry executed." }] })).resolves.toContain("accepted");
|
|
});
|