IchNukeDichWeg

Pygin — Development skill for Claude Code

Development community

Pygin - a from-scratch UCI chess engine in Python + C, with a hand-written search and evaluation and the entire per-node search loop in C. Built with the help of Claude Code.

How to install Pygin

This entry records only its repository, not the path inside it, so there is no exact command to give. Open IchNukeDichWeg/Pygin and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Pygin does

Pygin - a from-scratch UCI chess engine in Python + C, with a hand-written search and evaluation and the entire per-node search loop in C. Built with the help of Claude Code.

Alternatives in Development

Gopher Code — Claude Code, rewritten from scratch in Go 714 ★
Pi — PI Wisdom-in-Action Engine v23 — top mandates, scene router, delivery quality gates, anti-patterns, battle mom 107 ★
Tidewright — A 3D sandcastle simulator that runs entirely on the GPU 76 ★

README

Pygin

**A from-scratch chess engine in Python + C.** The search is hand-written. Since v58 the evaluation is an HCE/NNUE hybrid: the neural net scores positions inside the main search, and the hand-crafted eval keeps quiescence. There is no external engine, and the net learned from Pygin's own self-play games only, with no borrowed data and no borrowed weights.
[`python-chess`](https://pypi.org/project/chess/) is used *only* for board representation, move generation and legality.

![Strength](https://img.shields.io/badge/strength-~2868_Elo-3fb950) ![Speed](https://img.shields.io/badge/speed-3.3M_nps-58a6ff) ![Versions](https://img.shields.io/badge/versions-60-8b949e) ![C--era_gains](https://img.shields.io/badge/C--era_gains-%2B338_Elo-f0883e) ![Source](https://img.shields.io/badge/source-MIT-green) · Built with **[Claude Code](https://claude.com/claude-code)**

At a glance


~2868 Elo	SF-18 UCI_Elo scale	3.3M nps	the net costs ~30% of it
~+338 Elo	A/B-confirmed, v31→v60	~18 ply	from startpos in 5 s
+19.11 Elo	v58 NNUE, TIMED	1.43x	single-thread vs v31
v53+v54 eval lane	+37.52 & +31.20, the two biggest	1 dependency	`python-chess` only

Cumulative A/B Elo across the C era, v31=0 climbing to +324

Single-thread speed as a multiple of v31, peaking at 1.79x and ending at 1.43x once the net is armed

Both charts are self-play. Every C-era version (v31 and up) is A/B-tested against the one before it, and the gains stack to about +338 Elo. Single-thread speed peaked at 1.79× and sits at 1.43× today: v58 hands about 30% of it to the net and still comes out +19.11 ahead. The v30→v31 C rewrite (~34× faster) is off the left edge, so v31 is the honest zero.

The odds lad