mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-15 01:52:55 +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>
14 lines
4.6 KiB
TypeScript
14 lines
4.6 KiB
TypeScript
import { expect, it } from "vitest";
|
|
import type { EnvironmentLease } from "../../../packages/shared/src/types/environment.js";
|
|
import { destroyedLeaseBound, gradeLostSandbox, lostSandboxGuidance } from "./runtime-lost-sandbox.js";
|
|
function valid() { return { companyId: "company", leadId: "ceo", scenarioId: "workspaces.recover-lost-sandbox", origin: "http://127.0.0.1:42123", reply: "", links: [], agents: [], humanQueries: [], activity: [], tasks: [{ id: "website" }], runs: [{ id: "old", status: "failed" }, { id: "canonical", status: "succeeded" }, { id: "reconstruction", status: "succeeded" }, { id: "report", status: "succeeded" }], state: { task: { id: "website" }, exactProviderDestroyed: true, platformDestructionBound: true, survivingAttachmentVerified: true, noReconstructionBeforeHuman: true, guidanceReply: "The original sandbox was destroyed. A board member can use Continue from saved sources with the surviving attachments to reconstruct the page; missing unregistered files are not recovered.", actualReconstructionUi: true, actualReconstructionStatus: 202, humanDecisionBound: true, missingWorkAcknowledged: true, freshSessionBound: true, decisionSourcesBound: true, reconstructedArtifactVerified: true, newerWorkPreserved: true, originalFailureRetained: true, originalEventsRetained: true, confirmationBound: true, confirmationReply: "The website was reconstructed from the surviving attachment and the newer edits were preserved. The original run remains failed; missing unregistered state from the destroyed workspace is still unavailable." } }; }
|
|
it("accepts real authorized reconstruction while preserving the old failure", () => expect(gradeLostSandbox(valid()).passed).toBe(true));
|
|
it("accepts a natural saved-source guidance and factual result paraphrase", () => { const o = valid(); o.state.guidanceReply = "You need to choose the saved attachments and acknowledge the lost workspace state before rebuilding. An operator owns that decision."; o.state.confirmationReply = "Only the attached source was used to rebuild the page; your newer files were preserved separately. Unattached work cannot be restored from the destroyed allocation."; expect(gradeLostSandbox(o).passed).toBe(true); });
|
|
it("rejects SDK absence alone without independent platform evidence", () => { const o = valid(); o.state.platformDestructionBound = false; expect(gradeLostSandbox(o).passed).toBe(false); });
|
|
it("rejects an API-only substitute for the real routed human decision", () => { const o = valid(); o.state.actualReconstructionUi = false; expect(gradeLostSandbox(o).passed).toBe(false); });
|
|
it("rejects wrong decision source, stale session, missing newer edits and wrong reply phase", () => { for (const key of ["decisionSourcesBound", "freshSessionBound", "newerWorkPreserved", "confirmationBound", "humanDecisionBound"] as const) { const o = valid(); o.state[key] = false; expect(gradeLostSandbox(o).passed).toBe(false); } });
|
|
it("rejects erasing the failed source or claiming all lost files recovered", () => { const o = valid(); o.runs[0]!.status = "succeeded"; expect(gradeLostSandbox(o).passed).toBe(false); const p = valid(); p.state.confirmationReply += " All lost work was recovered."; expect(gradeLostSandbox(p).passed).toBe(false); expect(lostSandboxGuidance("I already reconstructed all lost files from that destroyed sandbox.")).toBe(false); });
|
|
it("requires the exact persisted destroyed termination receipt and no ongoing cleanup", () => { const binding = { companyId: "company", taskId: "website", runId: "old", leaseId: "lease", sandboxId: "sandbox", environmentId: "environment" }; const lease = { id: "lease", companyId: "company", issueId: "website", heartbeatRunId: "old", environmentId: "environment", provider: "daytona", providerLeaseId: "sandbox", metadata: { remoteExecutionTermination: { state: "destroyed" }, pendingCleanupInFlight: false } } as unknown as EnvironmentLease; expect(destroyedLeaseBound(lease, binding)).toBe(true); expect(destroyedLeaseBound({ ...lease, providerLeaseId: "another" }, binding)).toBe(false); expect(destroyedLeaseBound({ ...lease, metadata: { remoteExecutionTermination: { state: "stopped" } } }, binding)).toBe(false); expect(destroyedLeaseBound({ ...lease, metadata: { ...lease.metadata, pendingCleanupInFlight: true } }, binding)).toBe(false); });
|
|
|
|
it("rejects guidance that waives the actual human reconstruction decision", () => { expect(lostSandboxGuidance("The sandbox was destroyed; you can reconstruct from saved attachments without any board approval.")).toBe(false); expect(lostSandboxGuidance("You cannot rebuild the lost workspace without human approval of the saved attachments.")).toBe(true); });
|