Make your computer a home for your AI partner. Avibe is the local-first Agent OS: Vibey, its built-in AI partner, works out of the box, and Claude Code, Codex & OpenCode run unchanged beside it — reach them from your browser, phone, or any chat app.
Python agentagent-osaiai-agentsai-partnerchatopsclaudeclaude-codecodexdevtoolsdiscord-botlark-botllmlocal-firstlong-term-memoryopencodeslack-botvibe-codingvibeywechat
import { describe, expect, it } from 'vitest';
import {
chatTriggerLink,
harnessChipLabelKey,
needsHarnessProvenanceReconcile,
vaultCallbackStatusKey,
} from './chatTrigger';
type Msg = Parameters<typeof chatTriggerLink>[0];
const msg = (over: Partial<Msg>): Msg => ({
source: 'harness',
author_name: null,
author_id: null,
session_id: 'ses_here',
source_session_id: null,
source_session_title: null,
source_session_agent_name: null,
...over,
});
// Second arg is the localized "agent" fallback word; a fixed literal keeps the
// mapper translation-free and the test deterministic.
const link = (over: Partial<Msg>) => chatTriggerLink(msg(over), 'agent');
describe('chatTriggerLink', () => {
it('returns null for non-harness messages', () => {
expect(link({ source: 'user' })).toBeNull();
expect(link({ source: 'agent' })).toBeNull();
});
it('A9a: agent-callback links to the source session chat, labelled by title prefix', () => {
expect(link({ author_name: 'agent_run', source_session_id: 'ses_src_1234', source_session_title: 'Vaults 总控' })).toEqual({
kind: 'source',
to: '/chat/ses_src_1234',
label: 'Vaults 总控',
});
});
it('A9a: truncates a long source title to ~12 chars', () => {
expect(link({ source_session_id: 'ses_x', source_session_title: '0123456789ABCDEF' })).toEqual({
kind: 'source',
to: '/chat/ses_x',
label: '0123456789AB…',
});
});
it('A9a: falls back to agent name + short id when the title is null', () => {
expect(link({ source_session_id: 'ses_abcdef123456', source_session_title: null, source_session_agent_name: 'pm' })).toEqual({
kind: 'source',
to: '/chat/ses_abcdef123456',
label: 'pm · 123456',
});
});
it('A9a: uses the localized agent fallback when both title and agent name are null', () => {
expect(link({ source_session_id: 'ses_abcdef123456' })?.label).toBe('agent · 123456');
});
it('A9b: watch trigger links to the Harness watches tab filtered to this session', () => {
expect(link({ author_name: 'watch', author_id: 'def_w', session_id: 'ses_1' })).toEqual({
kind: 'harness',
to: '/harness?tab=watches&session=ses_1',
});
});
it('A9b: scheduled + task_run + legacy task all link to the Harness tasks tab', () => {
for (const author_name of ['scheduled', 'task_run', 'task']) {
expect(link({ author_name, session_id: 'ses_1' })?.to).toBe('/harness?tab=tasks&session=ses_1');
}
});
it('does not navigate for webhook or unknown harness kinds without a source', () => {
expect(link({ author_name: 'webhook' })).toBeNull();
expect(link({ author_name: 'show_intent' })).toBeNull();
expect(link({ author_name: 'agent_run', source_session_id: null })).toBeNull();
});
it('keeps an explicitly sourced watch on its exact source link', () => {
expect(link({ author_name: 'watch', source_session_id: 'ses_src' })?.kind).toBe('source');
});
});
describe('needsHarnessProvenanceReconcile', () => {
const m = (over: Partial<Parameters<typeof needsHarnessProvenanceReconcile>[0]>) =>
needsHarnessProvenanceReconcile({
source: 'harness',
native_message_id: 'agent_run:exec1',
source_session_id: null,
author_name: 'agent_run',
...over,
});
it('is true for a live agent_run harness message without a resolved source', () => {
expect(m({})).toBe(true);
});
it('is false once the source session is resolved', () => {
expect(m({ source_session_id: 'ses_src' })).toBe(false);
});
it('also reconciles legacy trigger rows whose native identity outlived provenance', () => {
for (const native_message_id of [
'scheduled:def:exec',
'watch:def:exec',
'webhook:def:exec',
'hook:exec',
]) {
expect(m({ native_message_id, author_name: null })).toBe(true);
}
});
it('is false for complete non-agent trigger rows and non-harness messages', () => {
expect(m({ native_message_id: 'scheduled:def:exec', author_name: 'scheduled' })).toBe(false);
expect(m({ native_message_id: null })).toBe(false);
expect(m({ source: 'agent' })).toBe(false);
});
});
describe('harnessChipLabelKey (leading-label key by source presence)', () => {
const key = (over: Partial<Msg>) => harnessChipLabelKey(msg(over));
it('a resolved agent-callback uses the "From" PREFIX key (precedes the source title)', () => {
expect(key({ author_name: 'agent_run', source_session_id: 'ses_src_1234' })).toBe('chat.source.from');
// The prefix wins even for a watch-authored row that carries a source.
expect(key({ author_name: 'watch', source_session_id: 'ses_src' })).toBe('chat.source.from');
});
it('an unresolved agent-callback falls back to the self-contained "From agent" key', () => {
// Source deleted / not yet enriched → no dangling bare "From".
expect(key({ author_name: 'agent_run', source_session_id: null })).toBe('chat.source.harness');
expect(key({ author_name: null, source_session_id: null })).toBe('chat.source.harness');
});
it('labels Vault callbacks separately from Agent callbacks', () => {
const metadata = {
source_kind: 'callback',
source_actor: 'vault:vrq_1',
vault_request_type: 'access',
vault_request_status: 'denied',
};
expect(key({ author_name: 'agent_run', metadata })).toBe('chat.source.vault');
expect(vaultCallbackStatusKey(msg({ author_name: 'agent_run', metadata }))).toBe('chat.source.vaultDenied');
});
it('uses the Vault outcome wording for provision, access, and sign callbacks', () => {
const callback = (requestType: string, status: string) =>
vaultCallbackStatusKey(msg({
metadata: {
source_kind: 'callback',
source_actor: 'vault:vrq_1',
vault_request_type: requestType,
vault_request_status: status,
},
}));
expect(callback('provision', 'fulfilled')).toBe('chat.source.vaultProvided');
expect(callback('access', 'approved')).toBe('chat.source.vaultAccessApproved');
expect(callback('sign', 'approved')).toBe('chat.source.vaultSigned');
expect(callback('delete', 'approved')).toBe('chat.source.vaultDeleteApproved');
expect(callback('delete', 'fulfilled')).toBe('chat.source.vaultDeleteApproved');
expect(callback('delete', 'denied')).toBe('chat.source.vaultDeleteDenied');
expect(callback('delete', 'expired')).toBe('chat.source.vaultDeleteExpired');
expect(callback('access', 'expired')).toBe('chat.source.vaultExpired');
});
it('every non-agent Harness trigger keeps its own self-contained label', () => {
expect(key({ author_name: 'watch' })).toBe('chat.source.watch');
expect(key({ author_name: 'webhook' })).toBe('chat.source.webhook');
expect(key({ author_name: 'hook' })).toBe('chat.source.hook');
expect(key({ author_name: 'activity_recovery' })).toBe('chat.source.activityRecovery');
for (const author_name of ['scheduled', 'task_run', 'task']) {
expect(key({ author_name })).toBe('chat.source.scheduled');
}
});
it('keeps the Show-specific label for a page-button intent', () => {
expect(key({ author_name: 'show_intent' })).toBe('chat.source.showIntent');
});
// An annotation is its own message type now and draws its own card, so it
// never reaches the harness chip. The branch that gave it a Show-specific
// label is gone along with the ``chat.source.showAnnotation`` string; this
// pins that it is not quietly re-added on the author_name side channel.
it('has no Show-specific label left for the retired show_annotation trigger', () => {
expect(key({ author_name: 'show_annotation' })).toBe('chat.source.harness');
});
});