HN · GitHub Signals

OPEN SOURCE, IN FOCUS

backpass

kunchenguid/backpass

You don't write AGENTS.md. You train it with gradient descent.

Explore on GitHub ↗

Saved projects stay in this browser. No account needed.

TAKE A CLOSER LOOK

Start with the source.

The repository is the place to check installation steps, supported platforms, the license, and current maintenance before choosing a tool.

Read the documentation on GitHub ↗

Keep exploring

Find projects in the same areas, or search the archive for a different approach.

ANOTHER GOOD FIND

Explore related projects

View the archive ↗

Connected by topics and descriptions.

hrtan

fastmix

[ICLR-2026] The PyTorch Implementation of FastMix. The algorithms are based on our paper: "Fast Data Mixture Optimization via Gradient Descent". FastMix is designed to rapidly identify optimal data mixing ratios for large language models (LLMs). - GitHub - hrtan/fastmix: [ICLR-2026] The PyTorch Implementation of FastMix. The algorithms are based on our paper: "Fast Data Mixture Optimization via Gradient Descent". FastMix is designed to rapidly identify optimal data mixing ratios for large language models (LLMs).

Take a look →
green-pt

honey-for-devs

Honey (I Shrunk the AI) by GreenPT: a cross-tool coding skill that cuts AI coding-agent token usage and LLM API costs — write less code, less prose, and denser agent-to-agent handoffs (−53%, lossless in benchmarks) with no loss of quality. Works with Claude Code, Cursor, GitHub Copilot, Codex, Gemini CLI, Windsurf, Cline & Kiro.

Take a look →
jetbrains

benjamin-plus-skill

Benjamin-Plus: a measured token-efficiency skill for coding agents (−17.9% cost median, quality unchanged). Inject it, don't install it.

Take a look →
001tmf

harness-forge

Turn Claude Code into its own Meta-Harness — a skill that evolves the scaffolding around a fixed model (memory, retrieval, context, prompts) via a native propose→score→Pareto loop. Native reimplementation of Meta-Harness (Lee et al. 2026).

Take a look →