sonmat (손맛) — Claude Code plugin that builds verification habits into AI-human collaboration
A protocol for agent work that survives session boundaries — plan, implement, verify in sealed contexts, with proof bundles as the only currency between them.
Discipline hooks for Claude Code. Stops the agent from claiming 'done' without proof, blocks accidental writes to ~/.claude config, and routes prompts to skills via local pgvector with zero Anthropic tokens.
A 200-line CLAUDE.md that fuses Andrej Karpathy's coding rules with field-tested agent-discipline patterns from real Claude Code builds.
Add a description, image, and links to the agent-discipline topic page so that developers can more easily learn about it.
To associate your repository with the agent-discipline topic, visit your repo's landing page and select "manage topics."