Self-hosted, Self-Improving, Federated Everything App — dev management, agent orchestration, persistent mind, digital twin, second brain, and creative pipeline from a single dashboard over Tailscale to port the user into the OS
JavaScript ai-agentsdashboarddeveloper-toolsdigital-twinquantified-selfsecond-brainself-hostedtailscale
import { describe, it, expect } from 'vitest';
import { canRunTaskOutputHookWithoutPayload, getTaskInputHook, getTaskOutputHook, isProgrammaticIoTaskType, resolveTaskHookType } from './taskTypeHooks.js';
describe('taskTypeHooks registry', () => {
it('resolves both hooks for layered-intelligence to callables', async () => {
const input = await getTaskInputHook('layered-intelligence');
const output = await getTaskOutputHook('layered-intelligence');
expect(typeof input).toBe('function');
expect(typeof output).toBe('function');
});
it('resolves both hooks for issue-watcher to callables', async () => {
expect(typeof await getTaskInputHook('issue-watcher')).toBe('function');
expect(typeof await getTaskOutputHook('issue-watcher')).toBe('function');
});
it('keeps pr-reviewer input owned by its screened preflight while sharing only the action hook', async () => {
expect(await getTaskInputHook('pr-reviewer')).toBeNull();
expect(typeof await getTaskOutputHook('pr-reviewer')).toBe('function');
});
it('resolves an input hook but no output hook for user-action-review', async () => {
// Input-only registration: the empty-ledger skip is the whole hook. The
// agent's deliverable (filed issues / queued tasks) has no output payload.
expect(typeof await getTaskInputHook('user-action-review')).toBe('function');
expect(await getTaskOutputHook('user-action-review')).toBeNull();
});
it('returns null for a task type with no registered hooks', async () => {
expect(await getTaskInputHook('security')).toBeNull();
expect(await getTaskOutputHook('security')).toBeNull();
expect(await getTaskInputHook('does-not-exist')).toBeNull();
expect(await getTaskOutputHook('does-not-exist')).toBeNull();
});
it('declares only payload-independent hooks safe for sentinel-less recovery', () => {
// No registered type is payload-independent today (quota-burn was, until it
// stopped being a scheduled task type entirely — see quotaBurnRunner.js).
// The predicate must still fail CLOSED for everything else.
expect(canRunTaskOutputHookWithoutPayload('layered-intelligence')).toBe(false);
expect(canRunTaskOutputHookWithoutPayload('issue-watcher')).toBe(false);
expect(canRunTaskOutputHookWithoutPayload('does-not-exist')).toBe(false);
});
});
describe('isProgrammaticIoTaskType (#2700)', () => {
it('recognizes a registered programmatic-I/O task type', () => {
expect(isProgrammaticIoTaskType('layered-intelligence')).toBe(true);
expect(isProgrammaticIoTaskType('issue-watcher')).toBe(true);
expect(isProgrammaticIoTaskType('pr-reviewer')).toBe(true);
});
it('rejects unregistered types, non-strings, and inherited Object keys', () => {
expect(isProgrammaticIoTaskType('ui')).toBe(false);
expect(isProgrammaticIoTaskType('')).toBe(false);
expect(isProgrammaticIoTaskType(undefined)).toBe(false);
expect(isProgrammaticIoTaskType(null)).toBe(false);
// A truthiness check on the registry object would let these through.
expect(isProgrammaticIoTaskType('constructor')).toBe(false);
expect(isProgrammaticIoTaskType('toString')).toBe(false);
});
});
describe('resolveTaskHookType', () => {
it('prefers the live scheduled type and supports archived task projections', () => {
expect(resolveTaskHookType({ taskType: 'internal', metadata: { analysisType: 'issue-watcher', taskAnalysisType: 'layered-intelligence' } }))
.toBe('issue-watcher');
expect(resolveTaskHookType({ taskType: 'internal', metadata: { taskAnalysisType: 'issue-watcher' } }))
.toBe('issue-watcher');
expect(resolveTaskHookType({ taskType: 'layered-intelligence', metadata: {} })).toBe('layered-intelligence');
});
});