agents-mdlisted

# AGENTS.md Skill Create or update AGENTS.md to help LLM coding agents work effectively in this repository. ## Purpose AGENTS.md is repository-specific guidance for LLM agents. Unlike README (for humans), AGENTS.md focuses on: - Exact commands with flags (`pytest -v` not "run tests") - Common gotchas with specifics ("Mock HOME needs both environ and Path.home patches") - Task-to-file mapping ("Add CLI command → src/cli.py") - Patterns with ✅/❌ code examples **Target audience:** LLM agents starting fresh **Quality:** Under 100 lines, actionable, specific **Location:** Repository root only ## Workflow ### Step 1: Determine Mode Check if `AGENTS.md` exists at repository root: - **CREATE:** Generate new AGENTS.md from scratch - **UPDATE:** Read existing file, add/update sections preserving user customizations ### Step 2: Invoke prompt-engineer Skill **REQUIRED:** Use Skill tool to invoke `prompt-engineer` skill. **Why:** AGENTS.md is for LLM agents. Apply prompt engineering principles: explicit instructions, important info first (primacy bias), specific over vague, examples over explanations. ### Step 3: Systematic Exploration Explore repository to gather content for AGENTS.md sections: **PRIORITY 1: Commands (document first)** - Check Makefile (`make help`), Justfile (`just --list`), package.json (scripts) - Read README.md Getting Started section - Check CI config (.github/workflows/) for canonical commands - Document: setup, build, run, test (full + file), lint, fo