What problem does it solve? Projects using multi-agent AI workflows often lack consistent AGENTS.md governance files, or their existing files drift out of compliance with operational rules like Zero-Autonomous-Commit, Kanban task management, and defensive shell standards. This Skill audits existing AGENTS.md files against a strict criteria checklist or generates compliant ones from scratch. ## Core Features & Use Cases - Phase 0 Generation: Creates AGENTS.md and docs/conventions.md from templates for new projects, filling placeholders with detected tech stack details. - Audit & Patch Mode: Reads existing AGENTS.md read-only first, evaluates it against 20+ criteria (Kanban directories, MCP tool usage, Gatekeeper halt protocol, Lite Mode, gitignore rules), then patches gaps while preserving formatting. - Standards Enforcement: Ships templates for architecture.md, DESIGN.md, and conventions.md covering the Universal DateTime Standard, SOLID guidelines, Financial Ledger Standard, and Defensive Shell Protocol. - Use Case: Onboard a new repository to an orchestrated agent workflow by generating a compliant AGENTS.md, or audit an existing project to detect missing rules like the mandatory end-of-task sequence. ## Quick Start Audit the AGENTS.md file in this project against the audit-agents criteria and patch any missing operational rules.
