Skip to content

Instantly share code, notes, and snippets.

View tomlupo's full-sized avatar

Tomasz Wilczega tomlupo

View GitHub Profile
@tomlupo
tomlupo / agent-orchestration-comparison.md
Last active February 28, 2026 13:53
Multi-Agent Orchestration Comparison: Custom Swarm vs Antfarm vs Claude SDK Loop

Multi-Agent Orchestration: Comparison Report

Comparing four approaches to AI agent coordination tested in February 2026.


1. Custom Swarm on OpenClaw (Discord-native)

What it is: Hand-configured multi-agent system using OpenClaw as the gateway. A coordinator agent (Atlas) receives natural language tasks via Discord, decomposes them, and delegates to specialized agents (Scout/research, Max/build, John/analytics, Maya/monitoring, Luna/content) using sessions_spawn. Results auto-announce back to Atlas for synthesis.

Typical Quant Investing Research Pipelines

1. Idea Generation / Alpha Research

  • Academic paper review (SSRN, NBER, Journal of Finance)
  • Anomaly identification (momentum, value, carry, mean-reversion)
  • Alternative data scouting (satellite, NLP sentiment, web scraping)
  • Factor decomposition of known strategies

2. Data Acquisition & Engineering

  • Market data: prices, volumes, order book (vendors: Bloomberg, Refinitiv, Polygon, Databento)
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>OpenClaw Memory System — Architecture + KSimback Integration</title>
<style>
:root{--bg:#0d1117;--card:#161b22;--border:#30363d;--text:#c9d1d9;--heading:#f0f6fc;--accent:#58a6ff;--green:#3fb950;--yellow:#d29922;--red:#f85149;--muted:#8b949e}
*{margin:0;padding:0;box-sizing:border-box}
body{font-family:-apple-system,BlinkMacSystemFont,'Segoe UI',Helvetica,Arial,sans-serif;background:var(--bg);color:var(--text);line-height:1.6;padding:2rem 1rem;max-width:820px;margin:0 auto}
{
"type": "excalidraw",
"version": 2,
"source": "claude-code",
"elements": [
{
"id": "title",
"type": "text",
"x": 280,
"y": 20,
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Speed Test</title>
<style>
:root{--bg:#0d1117;--card:#161b22;--border:#30363d;--text:#c9d1d9;--heading:#f0f6fc;--accent:#58a6ff;--green:#3fb950;--yellow:#d29922;--red:#f85149;--muted:#8b949e}
*{margin:0;padding:0;box-sizing:border-box}
body{font-family:-apple-system,BlinkMacSystemFont,'Segoe UI',Helvetica,Arial,sans-serif;background:var(--bg);color:var(--text);line-height:1.6;padding:2rem 1rem;max-width:820px;margin:0 auto}
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Give your OpenClaw the Memory it Needs — @KSimback Summary</title>
<style>
:root { --bg: #0d1117; --card: #161b22; --border: #30363d; --text: #c9d1d9; --heading: #f0f6fc; --accent: #58a6ff; --green: #3fb950; --yellow: #d29922; --red: #f85149; --muted: #8b949e; }
* { margin: 0; padding: 0; box-sizing: border-box; }
body { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Helvetica, Arial, sans-serif; background: var(--bg); color: var(--text); line-height: 1.6; padding: 2rem 1rem; max-width: 820px; margin: 0 auto; }
@tomlupo
tomlupo / pal_strategies_report.html
Created February 5, 2026 17:19
PAL Strategies Reverse Engineering Report
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>PAL Strategies Reverse Engineering Report</title>
<style>
:root {
--bg: #0d1117;
--card: #161b22;
@tomlupo
tomlupo / bankier-scrapes-comparison.md
Created February 4, 2026 08:29
Bankier Forum Scraper Repos – Full Comparison

Bankier Forum Scraper Repos – Full Comparison

Summary Table

Repo name Scraping technique Signal methodology Sentiment method Prediction generation Investing actionability Score (0–10) Recommendation Notes / Risks
ai-playground-claude-forum-rotation-signals httpx + BeautifulSoup; scrapes forum listing pages (forum_gielda) and parses thread titles; pagination by page index; retries on 503 with backoff; fixed delay; no headless browser Heuristic “rotation signal” scoring: keyword-based theme classification of thread titles + ticker extraction; composite score from mention volume, breadth, thread activity, sentiment bonus; outputs top themes and “high‑potential tickers” Polish lexicon (bullish/bearish word counts) over titles only Heuristic ranking (composite score + “HP score”); no predictive model Medium: surfaces themes/tickers, but no post content or price linkage 6 Improve Titles‑only sentiment miss
@tomlupo
tomlupo / ml_tactical_comparison.html
Created January 4, 2026 16:46
ML vs Tactical Model Comparison - Fund Rating System
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>ML vs Tactical Model Comparison | Fund Rating System</title>
<script src="https://cdn.jsdelivr.net/npm/chart.js"></script>
<style>
:root {
--primary: #2563eb;
<!DOCTYPE html>
<html>
<head>
<title>Tactical Pillar Model Rankings</title>
<style>
body { font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif; margin: 20px; background: #f5f5f5; }
h1 { color: #333; border-bottom: 2px solid #4CAF50; padding-bottom: 10px; }
h2 { color: #555; margin-top: 30px; }
h3 { color: #666; background: #e8e8e8; padding: 8px 12px; border-radius: 4px; }
.model-section { background: white; padding: 20px; margin: 20px 0; border-radius: 8px; box-shadow: 0 2px 4px rgba(0,0,0,0.1); }