d5bbce6dec
Three demo examples showcasing multi-agent orchestration: - **economics-seminar**: Hostile faculty panel debates AI economist presenter - **research-team**: Coordinator, Researcher, Analyst, Writer collaboration - **dungeon-master**: Persistent DM that creates its own game system 🤖 Generated with [Letta Code](https://letta.com) Co-Authored-By: Letta <noreply@letta.com>
130 lines
3.6 KiB
TypeScript
130 lines
3.6 KiB
TypeScript
#!/usr/bin/env bun
|
|
|
|
/**
|
|
* Economics Seminar CLI
|
|
*
|
|
* A multi-agent academic seminar simulation.
|
|
* An economist presents research, faculty panel asks questions.
|
|
*
|
|
* Usage:
|
|
* bun cli.ts # Run a seminar
|
|
* bun cli.ts --status # Show agent status
|
|
* bun cli.ts --reset # Reset all agents
|
|
* bun cli.ts --faculty=4 # Use 4 faculty members (default: 3)
|
|
* bun cli.ts --rounds=3 # Up to 3 Q&A rounds per faculty (default: 2)
|
|
*/
|
|
|
|
import { parseArgs } from 'node:util';
|
|
import * as readline from 'node:readline';
|
|
import { runSeminar, getStatus, resetSeminar } from './seminar.js';
|
|
import { DEFAULT_CONFIG } from './types.js';
|
|
|
|
/**
|
|
* Prompt user for input
|
|
*/
|
|
function prompt(question: string): Promise<string> {
|
|
const rl = readline.createInterface({
|
|
input: process.stdin,
|
|
output: process.stdout,
|
|
});
|
|
|
|
return new Promise((resolve) => {
|
|
rl.question(question, (answer) => {
|
|
rl.close();
|
|
resolve(answer.trim());
|
|
});
|
|
});
|
|
}
|
|
|
|
async function main() {
|
|
const { values } = parseArgs({
|
|
args: process.argv.slice(2),
|
|
options: {
|
|
status: { type: 'boolean', default: false },
|
|
reset: { type: 'boolean', default: false },
|
|
faculty: { type: 'string', default: '3' },
|
|
rounds: { type: 'string', default: '2' },
|
|
help: { type: 'boolean', short: 'h', default: false },
|
|
},
|
|
});
|
|
|
|
if (values.help) {
|
|
printHelp();
|
|
return;
|
|
}
|
|
|
|
if (values.status) {
|
|
await getStatus();
|
|
return;
|
|
}
|
|
|
|
if (values.reset) {
|
|
await resetSeminar();
|
|
return;
|
|
}
|
|
|
|
const config = {
|
|
...DEFAULT_CONFIG,
|
|
facultyCount: Math.min(4, Math.max(1, parseInt(values.faculty, 10) || 3)),
|
|
maxRoundsPerFaculty: Math.min(5, Math.max(1, parseInt(values.rounds, 10) || 2)),
|
|
};
|
|
|
|
// Prompt for topic
|
|
console.log('\n🎓 Economics Seminar\n');
|
|
console.log('Example topics:');
|
|
console.log(' - Impact of AI on labor markets');
|
|
console.log(' - Cryptocurrency regulation');
|
|
console.log(' - Universal basic income');
|
|
console.log(' - Housing affordability crisis');
|
|
console.log(' - Central bank digital currencies');
|
|
console.log('');
|
|
|
|
const topic = await prompt('📋 Enter a research topic (or press Enter for random): ');
|
|
|
|
await runSeminar(config, topic || undefined);
|
|
}
|
|
|
|
function printHelp() {
|
|
console.log(`
|
|
🎓 Economics Seminar
|
|
|
|
A multi-agent academic seminar simulation demonstrating:
|
|
- Agent collaboration and debate
|
|
- Persistent memory across sessions
|
|
- Distinct agent personalities
|
|
|
|
USAGE:
|
|
bun cli.ts [options]
|
|
|
|
OPTIONS:
|
|
--status Show current agent status and IDs
|
|
--reset Reset all agents (start fresh)
|
|
--faculty=N Number of faculty members (1-4, default: 3)
|
|
--rounds=N Max Q&A rounds per faculty (1-5, default: 2)
|
|
-h, --help Show this help
|
|
|
|
EXAMPLES:
|
|
bun cli.ts # Run with defaults (3 faculty, 2 rounds)
|
|
bun cli.ts --faculty=4 # Full panel of 4 faculty
|
|
bun cli.ts --rounds=1 # Quick seminar (1 question each)
|
|
bun cli.ts --status # See agent IDs and seminar count
|
|
|
|
THE SEMINAR:
|
|
1. 📋 You pick a topic (or let the presenter choose)
|
|
2. 📚 Presenter researches the topic using web search
|
|
3. 📖 Presenter gives their presentation
|
|
4. ❓ Hostile faculty panel attacks (back and forth)
|
|
5. 💭 Faculty delivers their brutal verdict
|
|
|
|
FACULTY PANEL:
|
|
👩🏫 Dr. Chen (Macro) - Policy implications, systemic effects
|
|
👨🏫 Dr. Roberts (Micro) - Incentives, equilibrium, theory
|
|
👩🏫 Dr. Patel (Behavioral) - Psychology, biases, real behavior
|
|
👴 Dr. Morrison (Historian) - Historical context, precedent
|
|
|
|
Each agent remembers past seminars and learns over time!
|
|
`);
|
|
}
|
|
|
|
main().catch(console.error);
|