Agent-optimized skills for the full LLM lifecycle — pre-training, post-training (RL/DPO/RLHF), inference, and autonomous research — plus GPU/TPU/QPU kernel programming, simulation, and scientific computing
2 files for Codex, OpenCode and Claude Code: agent-skills-for-compute AGENTS.md, agent-skills-for-compute CLAUDE.md — 698 tokens loaded in every session.
AGENTS.md A 261 tok CLAUDE.md A 437 tok These files are dtunai/agent-skills-for-compute's own configuration — they tell Codex, OpenCode and Claude Code how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.