What problem does it solve? AI coding assistants produce generic code that ignores a team's real conventions because nothing documents how the team works in a form the assistant can read. This Skill makes a repository agent-ready by writing its per-project config and rules layer, extracted from the actual codebase. ## Core Features & Use Cases - Convention Extraction: Maps the codebase and extracts its real conventions into coding rules under the configured standards location, citing real files and project idioms rather than generic templates. - Config Generation: Writes config_hints.json, AGENTS.md, .claude/skill.config, autonomous .claude/settings.json permissions, CLAUDE.md, PR/commit templates, and optional ERD documentation. - Stack Detection: Detects the project stack from positive evidence (e.g., Android vs Spring Boot vs generic JVM) and applies only the applicable framework rule directories. - Use Case: Point it at an unadopted repository, and it creates an install branch, explores the code, writes the full config and rules layer, verifies there is no contamination, and reports next steps. ## Quick Start Say "ar-install" or "make this repo agent-ready" and provide the full path to your target project when asked.
