Enforces strict orchestration rules and runtime validation for multi-agent workflows using the /team command.
Agent Enforcement provides a robust governance layer for multi-agent orchestration within Claude Code. It implements a two-layer defense strategy featuring a runtime PreToolUse hook to block orchestration violations and a model upgrade to Claude Opus for superior instruction following. By validating session directory paths, enforcing the use of specific agents like dev:researcher for internal tasks, and streamlining external model calls via the claudish CLI, this skill ensures reliable, secure, and deterministic behavior during complex team-based coding operations.
Características Principales
01Runtime violation blocking via PreToolUse hooks to prevent orchestration errors
02High-compliance instruction following optimized for Claude Opus
03Mandatory agent enforcement requiring dev:researcher for internal team tasks
04243 GitHub stars
05Deterministic claudish CLI integration for external model orchestration
06Secure session path validation favoring ai-docs/sessions over /tmp/
Casos de Uso
01Standardizing agent selection and task delegation across different model providers
02Debugging and preventing orchestration failures in complex multi-agent workflows
03Securing AI agent environments by restricting file path access during team tasks
