optim-agent v0.1.1

What's new

• Added ready-to-install plugins for Claude Code and Codex.
• Added public contribution, security, and project governance guidance.
• Added explicit development and optional vision dependency groups.
• Added sampler controls for trial-history length, explicit reasoning, and qualitative notes.

Improvements

• Separated lightweight core CI coverage from PyTorch vision tests.
• Refreshed benchmark documentation and public-facing experiment labels.
• Improved reproducibility and documentation for RL-control and credit-card benchmarks.
• Fixed Claude and Codex plugin manifest validation.

Removed

• Removed the contextual quant benchmark and its associated runner, figures, and artifacts.