Files
PaperClipAI/tests/runner-e2e/chat-readiness-observer.ts
T
f4337bb5fb test(evals): land the chat catalog eval harness on master (#19614)
## 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>
2026-11-10 08:02:11 -06:00

101 lines
7.5 KiB
TypeScript

import { retainHealthHelloDetail } from "./chat-readiness-detail.js";
import { constants } from "node:fs";
import { mkdir, open, readFile, realpath, stat } from "node:fs/promises";
import path from "node:path";
import type { ServerAdapterModule } from "../../server/src/adapters/types.js";
// Exact source-defined codes from OpenCode's existing environment test. Messages,
// hints, command output, config and credentials never cross this boundary.
const codes = new Set([
"opencode_environment_target", "opencode_cwd_valid", "opencode_cwd_invalid",
"opencode_openai_api_key_missing", "opencode_headless_permissions_enabled",
"opencode_command_skipped", "opencode_command_resolvable", "opencode_command_unresolvable",
"opencode_model_validation_via_probe", "opencode_model_invalid", "opencode_model_validation_skipped_remote",
"opencode_models_discovered", "opencode_models_empty", "opencode_models_discovery_failed",
"opencode_model_configured", "opencode_hello_probe_timed_out", "opencode_hello_probe_passed",
"opencode_hello_probe_unexpected_output", "opencode_hello_probe_model_unavailable",
"opencode_hello_probe_auth_required", "opencode_hello_probe_failed",
"paperclip_runner_provider_unsupported", "agent_runner_config_incompatible", "paperclip_runner_mode_invalid", "paperclip_runner_opencode_model_invalid",
]);
const categories = ["authentication", "rate-limit", "network", "permission", "command-missing", "module-missing", "model-unavailable", "timeout", "unknown"] as const;
type Category = typeof categories[number];
const uuid = (v: unknown): v is string => typeof v === "string" && /^[a-f0-9-]{36}$/i.test(v);
const selected = (ids: readonly string[]) => ids.length > 0 && ids.every(id => id === "chat-setup-health.runner-opencode.local.setup-only" || /^chat-features\.runner-opencode\.local\.[a-z0-9.-]+$/.test(id));
const filename = (root: string) => path.join(root, "artifacts-private", "chat-readiness-probes.jsonl");
export function readinessErrorCategory(value: unknown): Category {
const text = typeof value === "string" ? value.slice(0, 8192) : "";
if (/ProviderModelNotFoundError|model.{0,40}not found/i.test(text)) return "model-unavailable";
if (/\b(?:401|403|unauthorized|authentication|invalid.api.key)\b/i.test(text)) return "authentication";
if (/\b429\b|rate.?limit/i.test(text)) return "rate-limit";
if (/EACCES|EPERM|permission denied/i.test(text)) return "permission";
if (/ERR_MODULE_NOT_FOUND|Cannot find (?:module|package)/i.test(text)) return "module-missing";
if (/ENOENT|command not found/i.test(text)) return "command-missing";
if (/ETIMEDOUT|timed? ?out/i.test(text)) return "timeout";
if (/ECONNREFUSED|ENOTFOUND|ECONNRESET|fetch failed|TLS|certificate/i.test(text)) return "network";
return "unknown";
}
function closedCheck(value: any) {
return { code: codes.has(value?.code) ? value.code as string : "unknown",
level: ["info", "warn", "error"].includes(value?.level) ? value.level as string : "unknown",
category: categories.includes(value?.category) ? value.category as Category : "unknown" };
}
export async function installChatReadinessObserver(input: {
executionIds: readonly string[]; temporaryRoot: string; adapter: Pick<ServerAdapterModule, "testEnvironment">;
}) {
if (!selected(input.executionIds)) return;
const root = await realpath(input.temporaryRoot), info = await stat(root);
if (root !== input.temporaryRoot || !info.isDirectory() || info.uid !== process.getuid?.() || (info.mode & 0o077)) throw new Error("Readiness observer requires its owned private attempt root");
const directory = path.dirname(filename(root)); await mkdir(directory, { recursive: true, mode: 0o700 });
const directoryInfo = await stat(directory);
if (await realpath(directory) !== directory || directoryInfo.uid !== info.uid || (directoryInfo.mode & 0o022)) throw new Error("Readiness observer directory is not private and owned");
const original = input.adapter.testEnvironment;
let sequence = 0, writes = Promise.resolve();
const retain = async (companyId: string, result: any, thrown: boolean) => {
if (++sequence > 100) return;
const detailRetention = thrown ? undefined : await retainHealthHelloDetail(root, input.executionIds, sequence, result);
const receipt = { ...(detailRetention ? { detailRetention } : {}), schema: "paperclip.chat-readiness-probe/v1", companyId, sequence,
status: ["pass", "warn", "fail"].includes(result?.status) ? result.status : "unknown", thrown,
checks: Array.isArray(result?.checks) ? result.checks.slice(0, 64).map((check: any) => closedCheck({ ...check,
category: readinessErrorCategory(`${check?.message ?? ""}\n${check?.detail ?? ""}`) })) : [],
errorCategory: thrown ? readinessErrorCategory(result instanceof Error ? result.message : "") : "unknown" };
writes = writes.then(async () => {
const file = await open(filename(root), constants.O_WRONLY | constants.O_APPEND | constants.O_CREAT | constants.O_NOFOLLOW, 0o600);
try {
const fileInfo = await file.stat();
if (!fileInfo.isFile() || fileInfo.uid !== info.uid || (fileInfo.mode & 0o077)) throw new Error("Unsafe readiness receipt file");
await file.writeFile(`${JSON.stringify(receipt)}\n`);
} finally { await file.close(); }
}).catch(() => { /* Diagnostic failure must never replace the actual probe result/error. */ });
await writes;
};
input.adapter.testEnvironment = async function (...args) {
const context = args[0];
if (context.adapterType !== "paperclip_runner" || context.config.provider !== "opencode" || !uuid(context.companyId)) return original.apply(this, args);
let result;
try { result = await original.apply(this, args); }
catch (error) { await retain(context.companyId, error, true); throw error; }
await retain(context.companyId, result, false);
return result;
};
}
/** Re-project even the private file. No arbitrary file keys/strings enter evidence. */
export interface ClosedReadinessReceipt {
schema: "paperclip.chat-readiness-probe/v1"; companyId: string; sequence: number; status: string; thrown: boolean;
checks: Array<{ code: string; level: string; category: Category }>; errorCategory: Category; detailRetention?: "retained" | "withheld";
}
export async function readChatReadinessReceipts(temporaryRoot: string, companyId: string): Promise<ClosedReadinessReceipt[]> {
if (!temporaryRoot || !uuid(companyId)) return [];
let bytes: string;
try { bytes = await readFile(filename(temporaryRoot), "utf8"); }
catch (error) { if ((error as NodeJS.ErrnoException).code === "ENOENT") return []; throw new Error("Readiness receipt unavailable"); }
if (bytes.length > 1_000_000) throw new Error("Readiness receipt exceeds bound");
return bytes.trim().split("\n").filter(Boolean).slice(0, 100).flatMap(line => {
let row; try { row = JSON.parse(line); } catch { return []; }
if (!row || row.schema !== "paperclip.chat-readiness-probe/v1" || row.companyId !== companyId || !Number.isInteger(row.sequence) || row.sequence < 1 || row.sequence > 100) return [];
return [{ ...(row.detailRetention === "retained" || row.detailRetention === "withheld" ? { detailRetention: row.detailRetention } : {}), schema: row.schema, companyId, sequence: row.sequence,
status: ["pass", "warn", "fail"].includes(row.status) ? row.status : "unknown", thrown: row.thrown === true,
checks: Array.isArray(row.checks) ? row.checks.slice(0, 64).map(closedCheck) : [],
errorCategory: categories.includes(row.errorCategory) ? row.errorCategory as Category : "unknown" }];
});
}