Agent builds · leaderboard
Ranked by use, not hype.
Short answer: score = 3 per person who says they run it + 1 per star + 2 per fork + 2 per receipt point (a receipt counts 1, 2 if our checker found its number on the source, 4 if a moderator verified it).
| # | Build | Score | Use it |
|---|---|---|---|
| 1 | Site-first article and explainer agents @mattbusel · Claude Opus (Claude Code background agents) | 8 | 0 |
| 2 | AI outreach agent team: short-lived role agents on a schedule (free sample) @mattbusel · Claude Opus and Claude Sonnet (Claude Code background agents) | 8 | 0 |
| 3 | Weekly report skill (free sample from the AI Builder Kit) @mattbusel · Any Claude model that supports skills (Claude Code, Claude apps) | 0 | 0 |
| 4 | sqlswarm: a cheat-proof head-to-head for coding agents @mattbusel · Claude Opus 5.5 (Claude Code, both arms) | 0 | 0 |
| 5 | throng: a swarm of coding agents verified by real builds and tests @mattbusel · Local models (Ollama), the Anthropic API, or Claude Code workers | 0 | 0 |