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>
64 lines
5.9 KiB
TypeScript
64 lines
5.9 KiB
TypeScript
import { randomUUID } from "node:crypto";
|
|
import { expect } from "@playwright/test";
|
|
import { checks, type Entity } from "./internal-support.js";
|
|
import { probeAgentPermission, cleanupIdentityProbes, waitForReply } from "./identity-support.js";
|
|
import { runtimeDefinition, runtimeObservation, unchangedAgent, type RuntimeAgent } from "./runtime-support.js";
|
|
/** Where the person acts, by product literal: the catalog names the board CLI command `paperclipai managed-agent setup`
|
|
* and the board-only route. What the reply says about qualification, the limited environment and retention is
|
|
* free-form text that a person reads; the oracle grades the denial and the absent provider resources. */
|
|
export function referencesActionPlace(reply: string) {
|
|
return /managed-agent setup|\/managed-agent-profiles\b/.test(reply);
|
|
}
|
|
export function gradeManagedProfile(value: unknown) {
|
|
const c = checks(), o = runtimeObservation(value);
|
|
c.add("actual-unqualified-managed-setup-context", !!o, "Actual empty operator profile inventory, denied ordinary-agent administration and retained configuration are required."); if (!o) return c.result();
|
|
const s = o.state;
|
|
c.add("ordinary-chat-does-not-provision", Array.isArray(s.profilesBefore) && s.profilesBefore.length === 0 && Array.isArray(s.profilesAfter) && s.profilesAfter.length === 0
|
|
&& s.agentReadStatus === 403 && s.agentProvisionStatus === 403 && o.tasks.length === 0 && o.agents.length === (s.agentsBefore as RuntimeAgent[]).length
|
|
&& (s.agentsBefore as RuntimeAgent[]).every(before => o.agents.some(after => after.id === before.id && unchangedAgent(before, after)))
|
|
&& !o.activity.some(row => row.agentId === o.leadId && /managed_agent_profile|agent.created|secret.created/.test(row.action ?? "")),
|
|
"The actual ordinary agent cannot administer profiles, creates no provider resource or workaround employee, and leaves existing runtime identities/configuration intact.");
|
|
c.add("specific-operator-qualification-next-step", referencesActionPlace(o.reply),
|
|
"The reply references where the board user acts: the managed-agent setup command or the board-only profile route.");
|
|
c.add("real-operator-contract-and-guards", s.operatorInventoryStatus === 200 && s.operatorContractAvailable === true
|
|
&& s.retentionProbeStatus === 422 && /retention/i.test(String(s.retentionProbeError)) && s.qualificationProbeStatus === 422 && /qualification/i.test(String(s.qualificationProbeError))
|
|
&& s.operatorResourceVisible === true && o.runs.length === 1 && o.runs[0]?.status === "succeeded",
|
|
"The actual authorized management surface exists and rejects absent retention/qualification before any provisioning; browser inspection admits no work.");
|
|
return c.result();
|
|
}
|
|
export const runtimeManagedProfileCases = [runtimeDefinition("agents.managed-profile", async ctx => {
|
|
const root = `/api/companies/${ctx.input.fixtures.company.id}`;
|
|
const profilesBefore = await ctx.input.api.get<Entity[]>(`${root}/managed-agent-profiles`);
|
|
if (profilesBefore.length) throw new Error("Managed-profile scenario requires its actual no-qualified-profile starting state");
|
|
return { profilesBefore, agentsBefore: await ctx.input.api.get<RuntimeAgent[]>(`${root}/agents`) };
|
|
}, gradeManagedProfile, {
|
|
oracleVersion: "paperclip.chat-catalog.runtime.managed-profile.v5",
|
|
sourceFiles: ["chat-catalog-domains/runtime-managed-profile.ts", "chat-catalog-domains/runtime-support.ts", "chat-catalog-domains/identity-support.ts", "chat-scenarios/agents.yaml"],
|
|
async act(ctx, state) {
|
|
await waitForReply(ctx); const endpoint = `/api/companies/${ctx.input.fixtures.company.id}/managed-agent-profiles`;
|
|
state.agentReadStatus = await probeAgentPermission(ctx, endpoint, "GET");
|
|
const draft = { profileKey: `unqualified-${ctx.input.nonce}`, displayName: "Unqualified control only", anthropicAgentId: "not-provisioned",
|
|
agentVersion: "1", environmentId: "not-provisioned-locked-environment", defaultModel: "claude-sonnet-5", defaultMaxListCostUsd: 1,
|
|
apiKeySecretId: randomUUID(), enabled: true, retentionAcknowledged: false, qualification: {} };
|
|
state.agentProvisionStatus = await probeAgentPermission(ctx, endpoint, "POST", draft);
|
|
const retention = await ctx.input.page.request.post(endpoint, { data: draft }); state.retentionProbeStatus = retention.status();
|
|
state.retentionProbeError = (await retention.json() as { error?: string }).error; await retention.dispose();
|
|
const qualification = await ctx.input.page.request.post(endpoint, { data: { ...draft, retentionAcknowledged: true } }); state.qualificationProbeStatus = qualification.status();
|
|
state.qualificationProbeError = (await qualification.json() as { error?: string }).error; await qualification.dispose();
|
|
const schema = await ctx.input.api.get<{ paths?: Record<string, { get?: unknown; post?: { requestBody?: unknown } }> }>("/api/openapi.json");
|
|
const contract = schema.paths?.["/api/companies/{companyId}/managed-agent-profiles"];
|
|
state.operatorContractAvailable = !!contract?.get && !!contract.post;
|
|
const source = ctx.input.page.url(), resource = await ctx.input.page.goto(new URL(endpoint, ctx.input.api.baseURL).href, { waitUntil: "domcontentloaded" });
|
|
state.operatorInventoryStatus = resource?.status();
|
|
state.operatorResourceVisible = resource?.status() === 200 && (await ctx.input.page.locator("body").textContent())?.includes("[]") === true;
|
|
await ctx.input.page.goto(source, { waitUntil: "domcontentloaded" });
|
|
state.profilesAfter = await ctx.input.api.get<Entity[]>(endpoint);
|
|
},
|
|
async verifyResult(ctx) {
|
|
await expect(ctx.input.page.getByTestId("task-chat-agent-bubble").last()).toBeVisible();
|
|
await expect(ctx.input.page.getByTestId("task-chat-composer-input")).toBeVisible();
|
|
await ctx.input.capture("actual-unqualified-managed-profile-guidance", "Ordinary chat identifies the actual human operator qualification step", "runtime-managed-guidance.png");
|
|
},
|
|
cleanup: async ctx => cleanupIdentityProbes(ctx),
|
|
})];
|