Beautiful, AI-native markdown IDE and LLM wiki
TypeScript 2nd-brainagent-skillsclaudecodexcompany-braindocsknowledge-knowledge-baseknowledge-managementllmllm-wikillm-wiki-karpathymarkdownmarkdown-editormdmdxnotespkmsecond-brainskills
import { describe, expect, test } from 'vitest';
import { codexProbeCommand, parseCodexCatalog, probeCodexCatalog } from './codex-probe.ts';
const real = JSON.stringify({
models: [
{
slug: 'gpt-5.6-sol',
visibility: 'list',
context_window: 272_000,
max_context_window: 872_000,
effective_context_window_percent: 95,
},
{
slug: 'gpt-reserve',
visibility: 'hide',
context_window: 272_000,
max_context_window: 872_000,
effective_context_window_percent: 95,
},
],
});
describe('reading the Codex catalog', () => {
test('the shape the installed CLI emits is understood', () => {
expect(parseCodexCatalog(real)).toEqual([
{
slug: 'gpt-5.6-sol',
visibility: 'list',
contextWindow: 272_000,
maxContextWindow: 872_000,
effectivePercent: 95,
},
{
slug: 'gpt-reserve',
visibility: 'hide',
contextWindow: 272_000,
maxContextWindow: 872_000,
effectivePercent: 95,
},
]);
});
test('a bare array is accepted as well as a models wrapper', () => {
expect(parseCodexCatalog('[{"slug":"x"}]')).toEqual([
{
slug: 'x',
visibility: null,
contextWindow: null,
maxContextWindow: null,
effectivePercent: null,
},
]);
});
test('a row with no slug is dropped rather than rendered nameless', () => {
expect(parseCodexCatalog('{"models":[{"context_window":1},{"slug":""}]}')).toEqual([]);
});
test('missing numbers stay null instead of becoming zero', () => {
const [row] = parseCodexCatalog('{"models":[{"slug":"x","context_window":"lots"}]}');
expect(row?.contextWindow).toBeNull();
});
test('output that is not JSON yields no models rather than throwing', () => {
for (const raw of ['', 'not json', 'null', '{}', '{"models":"nope"}']) {
expect(parseCodexCatalog(raw)).toEqual([]);
}
});
});
describe('resolving the binary the launch path would use', () => {
test('a missing binary reads as not-found, not as a failure', async () => {
const probe = await probeCodexCatalog('definitely-not-a-real-binary-xyz', { PATH: '' });
expect(probe.outcome).toBe('not-found');
});
test('a non-zero exit is a failure, so the login-shell retry does not fire for it', async () => {
const probe = await probeCodexCatalog(process.execPath, process.env);
expect(probe.outcome).toBe('failed');
});
test('a windows .cmd shim runs through the shell wrap, not execed directly', () => {
const composed = codexProbeCommand('codex.cmd', { PATH: '' }, 'win32');
expect(composed.wrap).toBe(true);
expect(composed.cmd.toLowerCase()).toContain('cmd');
expect(composed.args.join(' ')).toContain('debug');
});
test('elsewhere the binary is invoked directly with the catalog subcommand', () => {
const composed = codexProbeCommand('codex', { PATH: '/usr/bin' }, 'darwin');
expect(composed.wrap).toBe(false);
expect(composed.cmd).toBe('codex');
expect(composed.args).toEqual(['debug', 'models']);
});
});