Files
PaperClipAI/tests/runner-e2e/evidence.ts
T
cceeb0aa66 test(runner): add everyday workflow evaluation harness (#13474)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work.
> - The native runner must support project work, delegation, hiring, and
service access.
> - Browser tests exposed lost connection access, rejected helper
events, and stalled recovery.
> - Some eval failures also came from incorrect fixtures and decision
controls.
> - This pull request fixes those paths and adds eight everyday workflow
stories.
> - The tests retain observed failures and verify delivered files
independently.
> - The benefit is repeatable evidence for common user tasks and their
remaining gaps.

## Linked Issues or Issue Description

Related work: #13404 contains earlier workflow fixes. #13300 and #13470
changed the CI contracts used by the harness security tests. Merged
companion:
[paperclip-evals#22](https://github.com/paperclipai/paperclip-evals/pull/22).

**What happened?**

Native ACPX sessions did not receive the assigned connection gateway.
Codex helper events could arrive before their spawn receipt and fail
thread validation. A parent continuation could take a shared workspace
before its child retried. A failed native continuation could leave the
task status without a clear recovery blocker. The eval harness also
confused tool approvals with new connection requests and could reject a
valid delegated download.

**Expected behavior**

Keep assigned gateway access and its approval checks. Verify helper
lineage before accepting helper progress. Let a waiting child proceed
before automatic parent recovery. Preserve a failed task's recovery
ownership. Grade the actual requested workflow and its delivered files.

**Steps to reproduce**

Run the everyday workflow suite with the native Codex and Claude
profiles. Exercise service approval, connection refusal, delegated
project work, and teammate reuse. The commands and case requirements are
in `tests/runner-e2e/EVERYDAY-WORKFLOWS.md`. Use `pnpm
test:runner-recovery` for controlled crash and replacement cases.

## What Changed

- Pass the scoped connection gateway binding through the native ACPX
host and sidecar.
- Recognize Codex helper lineage from parent metadata and spawn
receipts. Verify early helper events with `thread/read`. Keep helper
events separate from root completion authority.
- Guide agents to use persistent hiring, child tasks, dependency
records, and a blocked handoff while waiting for a child.
- Defer automatic parent recovery while a child has an active execution
path in the same shared workspace. Allow parent recovery when the child
needs review.
- Record Blocked status and recovery evidence when a failed native
continuation needs reconciliation, including existing active or
escalated incidents. Preserve their owner and retry budget.
- Add eight browser-driven workflow cases. Use real decision controls,
explicit child feedback delivery, managed hiring credentials, and
independent ZIP checks inside a bounded Docker sandbox. Verify sandbox
availability before task creation. Record screenshot SHA-256 at capture.
- Keep runner crash probes in controlled recovery tests. Preserve the
original failure when cleanup also fails.
- Display missing accounting and replay revisions as unavailable. Align
harness security assertions with the approved CI changes.
- Make the channel-rejection browser fixture bind its file after the
send captures its payload. This prevents live refresh from removing the
file before the simulated race.

## Verification

- Full workspace `pnpm -r typecheck` passed after merging current
master.
- Runner E2E typecheck passed. Harness unit tests passed: 216/216.
- Wake-queue database tests passed: 55/55. The two added
existing-incident tests failed before the fix and pass after it.
- Docker artifact calibration passed: 12/12. Host-file and host-loopback
isolation tests failed before the fix and pass after it. Read-only
delivery and output limits are also verified.
- Full `pnpm build` passed. Targeted recovery tests passed: 83/83.
- The channel-rejection browser test passed five consecutive runs after
fixing the fixture race found in CI.
- Local general-server (12,351 tests), UI (6,250), CLI (485), and
workspace package groups passed. The monolithic run stopped at an
unchanged lock-heartbeat fixture race; the isolated workspace group
passed on rerun (shared: 747/747). A separate local serialized run
passed 97 files before two socket errors in the unchanged issue-list
route suite; that suite passed 15/15 on isolated rerun. These local full
commands did not finish uninterrupted; the complete CI matrix below
covers the remaining suites.
- Final head `0fb293733fe307be7e6667ae8f1364077d0c6455`: **34 successful
checks, 2 expected skips**, including every server/workspace shard,
browser shard, native runner verification, build, and typecheck. [Final
CI
run](https://github.com/paperclipai/paperclip/actions/runs/34989136700).
- Greptile reviewed this exact head at **5/5**; all review threads are
resolved. Both Superagent security checks are successful.
- ACPX credential-boundary tests passed: 118/118. Superagent accepted
the runner/sidecar versus provider-environment trace and cleared its
finding.
- The latest paid local campaign on source
`f6a2fdf7ac2af859826a2ae627ff4125a5478529` passed 22/24 cases: Sol 8/8,
Claude 7/8, Mini 7/8. These results predate the merge with current
master.
- The two remaining failures are in `hire-reuse`: Claude exceeded the
attempt deadline during final review; Mini made invalid deliverable tool
calls and remained Blocked.
- Six Daytona cases were not run because the matching immutable runner
image was unavailable. This PR does not claim new remote model results.

## Risks

The changes affect connection admission, helper identity, and recovery
scheduling. Assigned gateway grants and user approval still govern
service calls. The workspace admission gate still exists; the broader
folder-sync design is separate work. Provider behavior can still cause
the two recorded hiring failures. No database migration is required.
Paid cases are opt-in and have bounded attempt deadlines. Project
stories now require Docker and the documented pinned Python image on the
harness host.

## Model Used

OpenAI `gpt-6-astra` performed implementation, diagnosis, and
substantive review. OpenAI `gpt-5.6-luna` assisted with verification, PR
preparation, and review tracking. Both used repository tools and code
execution. Context-window sizes were not recorded.

## 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 (focused checks and
isolated reruns; full-run limitations are documented above)
- [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
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge

---------

Co-authored-by: OpenAI GPT-5.6 Luna <noreply@openai.com>
Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-09-15 11:04:16 -05:00

237 lines
7.5 KiB
TypeScript

import { spawn } from "node:child_process";
import {
mkdir,
readFile,
readdir,
rm,
writeFile,
copyFile,
} from "node:fs/promises";
import path from "node:path";
import {
assertSecretFree,
findSecretLeak,
findSecretLeakInJsonValues,
redactText,
sanitizeJson,
} from "./redaction.js";
const TEXT_EXTENSIONS = new Set([
".json",
".xml",
".html",
".css",
".js",
".txt",
".log",
".md",
]);
const BINARY_EXTENSIONS = new Set([".png", ".webm"]);
const ACTIVE_CONTENT_EXTENSIONS = new Set([".svg"]);
const ALLOWED_DIRECTORIES = new Set([
"snapshots",
"playwright-output",
"blob-report",
"html-report",
]);
const ALLOWED_ROOT_FILES = new Set([
"result.json",
"final-state.png",
"failure.png",
"tool-review-pending.png",
"decision-pending.png",
"chat-plan-draft.png",
"chat-plan-revised.png",
"server.log",
"playwright.log",
"junit.xml",
]);
const REQUIRED_PASS_FILES = [
"final-state.png",
"result.json",
"junit.xml",
path.join("html-report", "index.html"),
path.join("snapshots", "fixtures.json"),
path.join("snapshots", "api-state.json"),
] as const;
export interface EvidencePackageResult {
files: string[];
leaks: Array<{ file: string; reason: string }>;
missing: string[];
}
async function walk(root: string, relative = ""): Promise<string[]> {
const directory = path.join(root, relative);
const entries = await readdir(directory, { withFileTypes: true }).catch(
() => [],
);
const files: string[] = [];
for (const entry of entries) {
const next = path.join(relative, entry.name);
if (entry.isDirectory()) files.push(...(await walk(root, next)));
else if (entry.isFile()) files.push(next);
}
return files;
}
function isAllowed(relative: string) {
const segments = relative.split(path.sep);
if (segments.length === 1)
return (
ALLOWED_ROOT_FILES.has(relative) ||
/^(?:plan|question)-[a-z0-9-]+\.png$/.test(relative)
);
return ALLOWED_DIRECTORIES.has(segments[0]);
}
async function inspectZip(source: string, secrets: readonly string[]) {
// Failure traces can exceed Node's child-process output buffer. Stream the
// expanded archive through the exact-value scanner with enough overlap to
// detect a credential split across stdout chunks, without retaining the
// archive in memory or weakening fail-closed evidence publication.
const overlap = Math.max(
256,
...secrets.map((secret) => Buffer.byteLength(secret, "utf8") + 16),
);
return new Promise<string | null>((resolve) => {
const unzip = spawn("unzip", ["-p", source], {
stdio: ["ignore", "pipe", "pipe"],
});
let carry = Buffer.alloc(0);
let leak: string | null = null;
let spawnError: Error | null = null;
unzip.stdout.on("data", (chunk: Buffer) => {
if (leak) return;
const data = Buffer.concat([carry, chunk]);
leak = findSecretLeak(data, secrets, { includeShapes: false });
carry = data.subarray(Math.max(0, data.length - overlap));
if (leak) unzip.kill();
});
// Drain diagnostics so a noisy unzip cannot block. Error text is withheld
// because archive paths and contents belong to private attempt evidence.
unzip.stderr.resume();
unzip.on("error", (error) => {
spawnError = error;
});
unzip.on("close", (code) => {
if (leak) return resolve(leak);
if (spawnError) {
return resolve(`zip could not be inspected: ${spawnError.message}`);
}
return resolve(
code === 0
? null
: `zip could not be inspected: unzip exited with code ${String(code)}`,
);
});
});
}
export async function packageEvidence(input: {
privateDir: string;
uploadDir: string;
secrets: readonly string[];
expectPassScreenshot: boolean;
}): Promise<EvidencePackageResult> {
await rm(input.uploadDir, { recursive: true, force: true });
await mkdir(input.uploadDir, { recursive: true });
const files: string[] = [];
const leaks: EvidencePackageResult["leaks"] = [];
const available = await walk(input.privateDir);
for (const relative of available.filter(isAllowed)) {
const source = path.join(input.privateDir, relative);
const extension = path.extname(relative).toLowerCase();
const destination = path.join(input.uploadDir, relative);
await mkdir(path.dirname(destination), { recursive: true });
if (ACTIVE_CONTENT_EXTENSIONS.has(extension)) {
// SVG can execute script when opened directly from an artifact. Keep the
// source in the disposable private attempt directory, but never admit it
// to the sanitized CI artifact.
continue;
} else if (TEXT_EXTENSIONS.has(extension)) {
const raw = await readFile(source, "utf8");
const parsed = extension === ".json" ? JSON.parse(raw) : null;
const leak =
extension === ".json"
? findSecretLeakInJsonValues(parsed, input.secrets)
: findSecretLeak(raw, input.secrets);
if (leak) leaks.push({ file: relative, reason: leak });
// Redacting an already serialized JSON string can change escape
// boundaries around shell commands. Sanitize parsed values instead so
// uploaded snapshots stay valid JSON.
const safe =
extension === ".json"
? `${JSON.stringify(sanitizeJson(parsed, input.secrets), null, 2)}\n`
: redactText(raw, input.secrets);
if (extension === ".json") {
const safeLeak = findSecretLeakInJsonValues(
JSON.parse(safe),
input.secrets,
);
if (safeLeak)
throw new Error(`Secret leak in ${relative}: ${safeLeak}`);
} else {
assertSecretFree(safe, input.secrets, relative);
}
await writeFile(destination, safe, "utf8");
files.push(relative);
} else if (extension === ".zip") {
const leak = await inspectZip(source, input.secrets);
if (leak) {
leaks.push({ file: relative, reason: leak });
continue;
}
await copyFile(source, destination);
files.push(relative);
} else if (BINARY_EXTENSIONS.has(extension)) {
// This raw-byte scan catches embedded plaintext credentials, but cannot
// inspect rendered pixels. Provider/UI raster and video files remain in
// the access-controlled CI artifact. The public publisher creates its
// own synthetic summary image from fixed labels and numeric/status data.
const raw = await readFile(source);
const leak = findSecretLeak(raw, input.secrets);
if (leak) {
leaks.push({ file: relative, reason: leak });
continue;
}
await copyFile(source, destination);
files.push(relative);
}
}
const missing = input.expectPassScreenshot
? [
...REQUIRED_PASS_FILES.filter((required) => !files.includes(required)),
...(files.some(
(file) =>
file.startsWith(`blob-report${path.sep}`) && file.endsWith(".zip"),
)
? []
: [path.join("blob-report", "*.zip")]),
]
: [];
const manifest = {
schema: "paperclip.runner-e2e.evidence/v1",
files: [...files].sort(),
leaks,
missing,
};
const manifestText = `${JSON.stringify(sanitizeJson(manifest, input.secrets), null, 2)}\n`;
const manifestLeak = findSecretLeakInJsonValues(
JSON.parse(manifestText),
input.secrets,
);
if (manifestLeak)
throw new Error(`Secret leak in evidence-manifest.json: ${manifestLeak}`);
await writeFile(
path.join(input.uploadDir, "evidence-manifest.json"),
manifestText,
"utf8",
);
files.push("evidence-manifest.json");
return { files, leaks, missing };
}