What problem does it solve? Setting up a repository with consistent AI coding agent configuration (CLAUDE.md, copilot-instructions, personas, invariants, skills) is manual and error-prone. This Skill automates onboarding by detecting a repo's stacks and agents, then generating a tailored .ai-badger/ directory plus each agent's discovery files. ## Core Features & Use Cases - Automated Detection: detect.py proposes stacks (with transitive requires), coding agents (claude/copilot/hermes/pi), source control platform, and build/test/lint commands from the target repo. - Deterministic Scaffolding: scaffold.py builds .ai-badger/ (config.json, manifest.json, CLAUDE.md, agents/, instructions/, invariants/, skills/) and writes agent discovery copies, preserving hand-authored files and keep-marker regions. - Validation & Recovery: validate.py checks the authored config against the JSON schema, and a structured error-recovery table guides fixes for common failures. - Use Case: Point it at a new or mature repository, confirm the proposed config, and get a complete, idempotent agent-instruction setup with provenance recorded in manifest.json. ## Quick Start Ask the agent to run welcome-ai-badger to scaffold this repository with ai-badger agent instructions and confirm the detected stacks.