Files
PaperClipAI/server/src/__tests__/unikraft-connection.test.ts
T
Devin Foley 059ae49cab feat(connections): add Unikraft Cloud connection (#17139)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work.
> - Apps connections give agents governed access to external services
through vault credentials, grants, action policies, the tool gateway,
and audit.
> - Unikraft Cloud runs unikernel micro-VMs, services, and volumes
across regional metros, and it offers a sandbox plugin for commands and
files.
> - Unikraft is not in the catalog. It has no hosted MCP server, so the
generic "Connect your own MCP server" flow cannot reach it.
> - Daytona (#16483) added a native REST pattern for this case: a
catalog generated from the provider's published OpenAPI specifications
and executed through the gateway.
> - This pull request adds Unikraft on that pattern and moves the shared
native-provider code into one registry, so the shared connection code
does not name each provider.
> - The benefit is governed agent access to all published Unikraft
actions, with the token kept in the vault and sent only to published
Unikraft hosts.

## Linked Issues or Issue Description

Refs #16483

**Problem or motivation**

Teams that run workloads on Unikraft Cloud want their agents to list
instances, read logs, deploy and scale services, and run commands in
sandboxes. Unikraft is not in the Apps catalog. Unikraft has no hosted
MCP server, so operators cannot connect it through the remote MCP flow.

**Proposed solution**

Add a native Unikraft connection. Generate its actions from Unikraft's
official OpenAPI specifications for the platform API and the sandbox
plugin. Pin the specifications to a reviewed commit. The operator
connects an API token. The token stays in the vault and goes only to the
five published metro hosts. Every action takes an optional metro.
Actions without a metro use the connection's default metro, which the
operator can change under Advanced.

**Alternatives considered**

A catalog-only AppDefinition was not possible because no hosted MCP
endpoint exists. Wrapping the `unikraft` CLI was rejected because it
would expose host filesystem behavior and CLI login state. A second copy
of the Daytona bridge was rejected; the shared transport and catalog
contract are now one module with a provider registry.

**Roadmap alignment**

This extends the existing Apps and Connections provider coverage. It
uses the existing credential, grant, policy, gateway, and audit model.
It does not overlap planned core work.

## What Changed

- Added `scripts/generate-unikraft-tools.mjs` and its tests. The
generator reads the platform and sandbox plugin specifications at commit
`5dd107ea1cd8716b83cb86a83ed2a6ee46f99b5b`, resolves shared parameters,
records source SHA-256 hashes, and writes
`server/src/services/unikraft/operations.generated.json`.
- The catalog has 105 actions: 89 platform operations and 16 sandbox
plugin operations. The server-sent audit event stream is recorded in the
inventory and is not advertised as an action.
- Classification: GET operations and sandbox reads and waits are reads.
Delete, stop, suspend, detach, unpin, and sandbox command execution,
signals, and stdin are destructive. All other operations are writes.
- Added `server/src/services/connectors/unikraft/client.ts` and its
generated catalog. It validates arguments against the generated schemas,
accepts only published metros, rejects path traversal, sends JSON bodies
on GET and DELETE as the specification requires, and verifies the token
with `GET /v1/users/quotas`. A sandbox wait that times out (documented
HTTP 408) returns a still-running result instead of a failure.
- Moved the bounded, credential-redacting HTTP sender and the catalog
contract helpers from `daytona.ts` to `native-rest-client.ts`. Added
`native-provider-catalogs.ts`. `tool-access.ts`, the
`tool-access/catalog` and `token-broker` modules,
`connectors/tool-policy.ts`, and `tool-gateway.ts` now use this registry
instead of Daytona-specific checks. Daytona behavior does not change.
- Added the Unikraft definition source, the generated definition, the
brand mark and manifest entry, and the permission review (`liveProof:
not-run`).
- Added `doc/connections/UNIKRAFT.md` with a sequence diagram,
endpoints, setup, actions, and limits. Updated the connections README,
the permission audit, and both capability maps.
- Tests: definition shape, database-backed gateway reads, writes,
deletes, and sandbox commands for personal and shared grants, default
and per-call metros, rejection of unpublished metros and traversal, Off
policies, revocation, stale-contract blocking, credential export denial,
schema validation for all actions, pinned-transport tests for GET and
DELETE bodies, and a browser setup check.

## Verification

- `pnpm exec vitest run packages/shared/src
server/src/__tests__/unikraft-connection.test.ts
server/src/__tests__/unikraft-http-boundary.test.ts
server/src/__tests__/daytona-connection.test.ts
server/src/__tests__/daytona-http-boundary.test.ts
server/src/__tests__/tool-access-service.test.ts
server/src/__tests__/tool-gateway-service.test.ts
server/src/__tests__/tool-gateway.test.ts
ui/src/pages/apps/AppsConnect.test.tsx ui/src/pages/apps/Browse.test.tsx
ui/src/lib/app-brand-assets.test.ts` plus `connector-tool-policy` and
`tool-catalog-publication`: all passed after the rebase onto the split
tool-access modules.
- `node --test scripts/generate-unikraft-tools.test.mjs
scripts/generate-daytona-tools.test.mjs
scripts/app-brand-validation.test.mjs`: 20 passed.
- `node scripts/check-app-brand-assets.mjs`, `node
scripts/check-maps.mjs`, and `git diff --check`: passed.
- `pnpm --filter @paperclipai/shared typecheck`, `pnpm --filter
@paperclipai/server typecheck`, and `pnpm --filter @paperclipai/ui
typecheck`: clean.
- `npx playwright test --config tests/e2e/playwright.config.ts
tests/e2e/unikraft-connector.spec.ts
tests/e2e/daytona-connector.spec.ts`: 2 passed.
- `node scripts/generate-unikraft-tools.mjs` regenerates the committed
catalog with the recorded hashes.
- Manual: on a local instance, open Connectors, search for Unikraft, and
select **Add key**. Confirm the icon, the password field, the key
guidance, and the default metro select under **Change**.
- Not verified: a real Unikraft account. Live provider read and write
proof is pending.

## Risks

- A Unikraft token acts for the whole organization in every metro.
Unikraft offers no narrower token permissions. Assign only trusted
agents and review action permissions.
- Sandbox actions run arbitrary commands inside the selected instance.
They are classified as destructive.
- A timed-out write can still complete at the provider. The bridge does
not retry writes automatically.
- New metros fail closed until the specification is regenerated and
reviewed. Legacy `*.kraft.cloud` hosts are not reachable.
- The Daytona refactor moves code without behavior changes. All Daytona
tests pass without edits.
- The generated JSON snapshot is large because it holds the provider
input contracts. No migration or lockfile change is required.

## Model Used

- Claude Opus 5.5 (Anthropic), model ID `claude-opus-5-5`, 1M context
window, used through Claude Code with tool use (shell, file editing, web
research, browser tests).

## 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
2026-10-22 15:14:26 -07:00

199 lines
15 KiB
TypeScript

import { randomUUID } from "node:crypto";
import { and, eq } from "drizzle-orm";
import { afterAll, beforeAll, describe, expect, it, vi } from "vitest";
import { agents, companies, companyMemberships, companySecrets, connectionGrants, createDb,
ensurePostgresDatabase, applyPendingMigrations, heartbeatRuns, issues, toolCatalogEntries, toolConnections } from "@paperclipai/db";
import { toolAccessService } from "../services/tool-access.js";
import { createToolGatewayService } from "../services/tool-gateway.js";
import { UNIKRAFT_TOOLS, executeUnikraftTool, unikraftClient } from "../services/connectors/unikraft/client.js";
import { getEmbeddedPostgresTestSupport, startEmbeddedPostgresTestDatabase } from "./helpers/embedded-postgres.js";
const external = process.env.PAPERCLIP_TEST_DATABASE_URL;
const support = external ? { supported: true } : await getEmbeddedPostgresTestSupport();
const actor = { actorType: "user" as const, actorId: "unikraft-reviewer" };
const credential = "unikraft-fixture-token";
(support.supported ? describe : describe.skip)("Unikraft governed provider actions", () => {
let db: ReturnType<typeof createDb>;
let temp: Awaited<ReturnType<typeof startEmbeddedPostgresTestDatabase>> | undefined;
beforeAll(async () => {
if (external) {
const url = new URL(external); url.pathname = "/unikraft_" + randomUUID().replaceAll("-", "");
await ensurePostgresDatabase(external, url.pathname.slice(1));
await applyPendingMigrations(url.toString()); db = createDb(url.toString());
} else { temp = await startEmbeddedPostgresTestDatabase("unikraft-connector-"); db = createDb(temp.connectionString); }
}, 30000);
afterAll(async () => { await db?.$client.end(); await temp?.cleanup(); });
async function fixture(options: { personal?: boolean; metro?: string } = {}) {
const [company] = await db.insert(companies).values({ name: "Unikraft fixture", issuePrefix: `U${randomUUID().slice(0, 6)}` }).returning();
await db.insert(companyMemberships).values({ companyId: company.id, principalType: "user", principalId: actor.actorId, membershipRole: "admin", status: "active" });
const instanceId = randomUUID();
const request = vi.fn(async (url: string, init: RequestInit) => {
expect(new Headers(init.headers).get("Authorization")).toBe(`Bearer ${credential}`);
expect(["error", "manual"]).toContain(init.redirect);
const parsed = new URL(url);
expect(parsed.hostname).toMatch(/^api\.(dal|fra|sfo|sin|was)\.unikraft\.cloud$/);
const metro = parsed.hostname.split(".")[1];
const body = typeof init.body === "string" ? JSON.parse(init.body) : undefined;
if (parsed.pathname === "/v1/users/quotas") return Response.json({ status: "success", data: { quotas: [{ uuid: randomUUID() }] }, op_time_us: 1 });
if (parsed.pathname === "/v1/instances" && init.method === "GET") {
return Response.json({ status: "success", data: { instances: [{ uuid: instanceId, name: "web", metro, selector: body ?? null }] }, op_time_us: 1 });
}
if (parsed.pathname === "/v1/instances" && init.method === "POST") return Response.json({ status: "success", data: { instances: [{ uuid: instanceId, ...body }] }, op_time_us: 1 });
if (parsed.pathname === `/v1/instances/${instanceId}` && init.method === "DELETE") return Response.json({ status: "success", data: { instances: [{ uuid: instanceId, state: "deleted" }] }, op_time_us: 1 });
if (parsed.pathname === "/v1/instances/web/plugins/sandbox/commands" && init.method === "POST") {
return Response.json({ uuid: randomUUID(), cmd: body.cmd, stdout: `ran with ${credential}` });
}
throw new Error(`Unexpected provider request ${init.method} ${url}`);
});
const service = toolAccessService(db, { remoteHttpRequest: request, remoteHttpEndpointLookup: async () => [{ address: "8.8.8.8", family: 4 }] });
const connected = await service.connectGalleryApp(company.id, { galleryKey: "unikraft", methodKey: "api-token",
...(options.personal ? { grantKind: "user" as const } : {}), ...(options.metro ? { configValues: { metro: options.metro } } : {}),
credentialValues: { "credentials.authorization": credential } }, actor);
const [agent] = await db.insert(agents).values({ companyId: company.id, name: "Operator", role: "engineer", adapterType: "process" }).returning();
await service.finishGalleryAppConnection(company.id, connected.connectionId, { enabledCatalogEntryIds: connected.catalog.map(c => c.id),
askFirstCatalogEntryIds: [], access: { agentIds: [agent.id] } }, actor);
const [issue] = await db.insert(issues).values({ companyId: company.id, title: "Provider work", assigneeAgentId: agent.id }).returning();
const [run] = await db.insert(heartbeatRuns).values({ companyId: company.id, agentId: agent.id, issueId: issue.id,
scopeKind: "issue", invocationSource: "on_demand", status: "running", responsibleUserId: actor.actorId, contextSnapshot: { issueId: issue.id } }).returning();
const gateway = createToolGatewayService(db, { remoteHttpRequest: request });
const session = await gateway.createSession({ companyId: company.id, agentId: agent.id, runId: run.id, issueId: issue.id });
const tools = await gateway.listToolsForSession(session.token);
const call = (name: string, parameters: unknown) => {
const tool = tools.find(t => t.upstreamToolName === name);
expect(tool, name).toBeTruthy();
return gateway.executeTool({ sessionToken: session.token, tool: tool!.name, parameters, idempotencyKey: randomUUID() });
};
return { company, service, connected, agent, request, call, gateway, session, instanceId };
}
it.each([false, true])("uses the selected vault identity for real gateway reads, writes and sandbox commands (personal=%s)", async personal => {
const f = await fixture({ personal });
const [grant] = await db.select().from(connectionGrants).where(eq(connectionGrants.connectionId, f.connected.connectionId));
const [secret] = await db.select().from(companySecrets).where(eq(companySecrets.id, grant.credentialSecretRefs[0].secretId));
expect(grant.kind).toBe(personal ? "user" : "organization");
expect(secret.scope).toBe(personal ? "user" : "company");
expect(secret.ownerUserId).toBe(personal ? actor.actorId : null);
expect(f.connected.catalog).toHaveLength(UNIKRAFT_TOOLS.length);
expect(f.connected.catalog.find(c => c.toolName === "unikraft_get_instances")).toMatchObject({ isReadOnly: true });
expect(f.connected.catalog.find(c => c.toolName === "unikraft_create_instance")).toMatchObject({ isReadOnly: false, isDestructive: false });
for (const name of ["unikraft_delete_instance_by_uuid", "unikraft_stop_instances", "unikraft_sandbox_run_command"]) {
expect(f.connected.catalog.find(c => c.toolName === name), name).toMatchObject({ isDestructive: true, isReadOnly: false });
}
expect(JSON.stringify(await f.service.getConnection(f.connected.connectionId))).not.toContain(credential);
// The token is verified against the default metro.
expect(f.request.mock.calls.some(([url]) => url === "https://api.fra.unikraft.cloud/v1/users/quotas")).toBe(true);
// Batch selectors travel as a JSON body even on GET.
const listed = JSON.stringify(await f.call("unikraft_get_instances", { body: [{ name: "web" }] }));
expect(listed).toContain(f.instanceId); expect(listed).toContain('"metro":"fra"'); expect(listed).toContain('"name":"web"');
expect(JSON.stringify(await f.call("unikraft_get_instances", { metro: "sin" }))).toContain('"metro":"sin"');
expect(JSON.stringify(await f.call("unikraft_create_instance", { body: { image: "nginx:latest" } }))).toContain("nginx:latest");
expect(JSON.stringify(await f.call("unikraft_delete_instance_by_uuid", { path: { uuid: f.instanceId } }))).toContain("deleted");
const ran = JSON.stringify(await f.call("unikraft_sandbox_run_command", { instance: "web", body: { cmd: "echo done" } }));
expect(ran).toContain("echo done"); expect(ran).not.toContain(credential);
}, 30000);
it("uses the connection's default metro for verification and actions that do not name one", async () => {
const f = await fixture({ metro: "was" });
expect(f.request.mock.calls.some(([url]) => url === "https://api.was.unikraft.cloud/v1/users/quotas")).toBe(true);
expect(JSON.stringify(await f.call("unikraft_get_instances", {}))).toContain('"metro":"was"');
});
it("rejects unpublished metros, unknown arguments, path traversal and foreign bindings before forwarding the token", async () => {
const f = await fixture(); f.request.mockClear();
await expect(f.call("unikraft_get_instances", { metro: "attacker.example" })).rejects.toThrow();
await expect(f.call("unikraft_get_instances", { command: "printenv" })).rejects.toThrow();
await expect(f.call("unikraft_sandbox_run_command", { instance: "../../users/quotas", body: { cmd: "true" } })).rejects.toThrow("Invalid Unikraft path");
expect(f.request).not.toHaveBeenCalled();
await expect(executeUnikraftTool({ binding: { companyId: randomUUID(), agentId: f.agent.id },
connection: { companyId: f.company.id, config: {} }, headers: { Authorization: `Bearer ${credential}` }, request: f.request,
toolName: "unikraft_get_instances", parameters: {} })).rejects.toThrow("Connection not found");
await expect(executeUnikraftTool({ binding: { companyId: f.company.id, agentId: f.agent.id },
connection: { companyId: f.company.id, config: { methodConfig: { metro: "evil" } } }, headers: { Authorization: `Bearer ${credential}` },
request: f.request, toolName: "unikraft_get_instances", parameters: {} })).rejects.toThrow("Choose a Unikraft Cloud metro");
expect(f.request).not.toHaveBeenCalled();
});
it("rechecks revocation and Off policies for a previously created gateway session", async () => {
const f = await fixture();
await f.service.finishGalleryAppConnection(f.company.id, f.connected.connectionId, {
enabledCatalogEntryIds: f.connected.catalog.filter(c => c.toolName !== "unikraft_sandbox_run_command").map(c => c.id),
askFirstCatalogEntryIds: [], access: { agentIds: [f.agent.id] },
}, actor);
f.request.mockClear();
await expect(f.call("unikraft_sandbox_run_command", { instance: "web", body: { cmd: "true" } })).rejects.toThrow();
expect(f.request).not.toHaveBeenCalled();
await db.update(toolConnections).set({ enabled: false }).where(and(eq(toolConnections.id, f.connected.connectionId), eq(toolConnections.companyId, f.company.id)));
await expect(f.call("unikraft_get_instances", {})).rejects.toThrow();
expect(f.request).not.toHaveBeenCalled();
});
it("blocks a stale same-name contract and quarantines it on refresh", async () => {
const f = await fixture();
const changed = f.connected.catalog.find(entry => entry.toolName === "unikraft_sandbox_run_command")!;
await db.update(toolCatalogEntries).set({ versionHash: "earlier-provider-contract", schemaHash: "earlier-schema", riskLevel: "read" })
.where(eq(toolCatalogEntries.id, changed.id));
f.request.mockClear();
expect((await f.gateway.listToolsForSession(f.session.token)).some(tool => tool.upstreamToolName === changed.toolName)).toBe(false);
await expect(f.call(changed.toolName, { instance: "web", body: { cmd: "true" } })).rejects.toThrow();
expect(f.request).not.toHaveBeenCalled();
const refreshed = await f.service.refreshCatalog(f.connected.connectionId, actor, { enableAllByDefault: true });
expect(refreshed.catalog.find(entry => entry.toolName === changed.toolName)?.status).toBe("quarantined");
});
it("does not export the token to agents", async () => {
const f = await fixture();
const [run] = await db.select().from(heartbeatRuns).where(eq(heartbeatRuns.agentId, f.agent.id));
await expect(f.service.mintConnectionTokenForAgent({ connectionId: f.connected.connectionId, companyId: f.company.id,
agentId: f.agent.id, runId: run.id, body: {} as never })).rejects.toThrow("Unikraft credentials stay in the governed tool gateway");
});
});
describe("Unikraft provider client", () => {
it("validates every published action schema before provider dispatch", async () => {
const request = vi.fn(async () => { throw new Error("Unexpected provider request"); });
expect(UNIKRAFT_TOOLS).toHaveLength(105);
for (const tool of UNIKRAFT_TOOLS) {
await expect(executeUnikraftTool({ binding: { companyId: "company", agentId: "agent" },
connection: { companyId: "company", config: {} }, headers: { Authorization: `Bearer ${credential}` },
request, toolName: tool.name, parameters: null }), tool.name).rejects.toThrow("Invalid Unikraft action arguments");
}
expect(request).not.toHaveBeenCalled();
});
it("sanitizes provider denials and rejects an unconfirmed token", async () => {
await expect(unikraftClient({ Authorization: `Bearer ${credential}` }, async () => new Response(credential, { status: 401 })).probe())
.rejects.toThrow("Unikraft denied access");
await expect(unikraftClient({ Authorization: `Bearer ${credential}` }, async () => Response.json({ status: "error" })).probe())
.rejects.toThrow("did not confirm");
});
it("accepts the documented -1 drain timeout for the longest drain", async () => {
const request = vi.fn(async (_url: string, init: RequestInit) => Response.json({ status: "success", echo: JSON.parse(init.body as string) }));
for (const [toolName, parameters] of [
["unikraft_stop_instances", { body: [{ name: "web", drain_timeout_ms: -1 }] }],
["unikraft_suspend_instance_by_uuid", { path: { uuid: randomUUID() }, body: { drain_timeout_ms: -1 } }],
] as const) {
await executeUnikraftTool({ binding: { companyId: "company", agentId: "agent" }, connection: { companyId: "company", config: {} },
headers: { Authorization: `Bearer ${credential}` }, request, toolName, parameters });
}
expect(request).toHaveBeenCalledTimes(2);
await expect(executeUnikraftTool({ binding: { companyId: "company", agentId: "agent" }, connection: { companyId: "company", config: {} },
headers: { Authorization: `Bearer ${credential}` }, request, toolName: "unikraft_stop_instances",
parameters: { body: [{ name: "web", drain_timeout_ms: -2 }] } })).rejects.toThrow("Invalid Unikraft action arguments");
});
it("reports a timed-out sandbox wait as a still-running command, not a provider failure", async () => {
const commandId = randomUUID();
const request = vi.fn(async () => new Response(null, { status: 408 }));
const call = (toolName: string) => executeUnikraftTool({ binding: { companyId: "company", agentId: "agent" },
connection: { companyId: "company", config: {} }, headers: { Authorization: `Bearer ${credential}` }, request, toolName,
parameters: { instance: "web", path: { uuid: commandId }, ...(toolName.endsWith("with_timeout") ? { body: { timeout_s: 0.01 } } : {}) } });
await expect(call("unikraft_sandbox_wait_for_command_with_timeout")).resolves.toMatchObject({ status: "running", timedOut: true });
expect(request.mock.calls[0]![0]).toBe(`https://api.fra.unikraft.cloud/v1/instances/web/plugins/sandbox/commands/${commandId}/wait_timeout`);
// Other operations keep treating 408 as a failure.
await expect(call("unikraft_sandbox_get_command_by_uuid")).rejects.toThrow("Unikraft could not complete this operation");
});
it("bounds response bytes", async () => {
await expect(unikraftClient({ Authorization: `Bearer ${credential}` }, async () => new Response("x".repeat(8 * 1024 * 1024 + 1))).probe())
.rejects.toThrow("exceeded the limit");
});
});