Agentic Coding

Config files for LLM coding assistants. .agents/ and AGENTS.md are the source of truth; tool-specific locations symlink to them.

Structure

agentic-coding/
├── .agents/                    # Canonical config (edit here)
│   ├── settings.json           # Plugins, permissions
│   ├── commands/               # Slash commands (/command-name)
│   │   ├── ai-refactor.md
│   │   ├── commit-push.md
│   │   └── interview-me.md
│   ├── skills/                 # Domain knowledge
│   │   ├── agent-browser/
│   │   ├── coding-guidelines/
│   │   ├── justfile/
│   │   ├── value-realization/
│   │   ├── skill-creator/
│   │   ├── systematic-debugging/
│   │   ├── ruff/
│   │   └── uv/
│   ├── agents/                 # Custom agents (empty)
│   ├── rules/                  # Behavior rules (empty)
│   ├── hooks/                  # Hook scripts (empty)
│   └── workflows -> commands   # Compatibility alias
│
├── .claude/                    # Claude compatibility symlink
│   └── -> .agents
├── .codex/                     # OpenAI Codex
│   ├── prompts -> ../.agents/commands
│   ├── rules -> ../.agents/rules
│   └── skills -> ../.agents/skills
│
├── AGENTS.md                   # Canonical project instructions
└── CLAUDE.md -> AGENTS.md      # Claude compatibility alias

Quick Reference

Commands	Skills
/ai-refactor - code review	coding-guidelines
/commit-push - git workflow	systematic-debugging
/interview-me - spec refinement	skill-creator, ruff, uv

Extending

Add to .agents/:

• commands/*.md - new slash commands
• skills/<name>/SKILL.md - new skills
• agents/*.md - agent definitions
• rules/*.md - behavior constraints
• hooks/*.sh - lifecycle scripts

See AGENTS.md for format details.