Six agent harnesses compared: 120 lines vs 1.4M lines, feature by feature
PawelHuryn · x · 2026-11-05
A uniform LOC and feature comparison of six agent harnesses, verified against source/docs: Mini Harness (1,200 lines, 2 deps) is the only small one covering sessions, compaction, Claude's server-side compaction, permissions, subagents, goal mode, MCP and skills. mini-swe-agent (4,000 lines) is bash-only; nanocode is 220 lines; Pi is 115K lines of TS; Codex CLI is 1.4M lines of Rust with 77 deps. Notably Claude Code's own compaction drops thinking, while Mini Harness preserves it.
More from coding & agent
- tw93 open-sources Kami: a design system that makes AI-agent documents actually readable — sujingshen · 2026-11-05
- Power user shares the one thing he won't let Claude do: write his tweets — PrajwalTomar_ · 2026-11-05
- Old iPad reborn as a smart home wall panel, thanks to Claude — PrajwalTomar_ · 2026-11-05
- He replaced paid app subscriptions with Claude-built tools, saving $200/month — PrajwalTomar_ · 2026-11-05
- Running Claude Code on DeepSeek for chores: one job costs $0.0043 — PrajwalTomar_ · 2026-11-05
- Agent audited a 300K-star skill pack in 70 seconds, found startup hook and usage counter — PrajwalTomar_ · 2026-11-05