KS
Killer-Skills

root — how to use root how to use root, root setup guide, what is root, root alternative, root vs Hikaze-Model-Manager-1, root install, Codex project governance, Hikaze-Model-Manager-2 integration, root skill for AI agents, governing Codex projects with root

v1.0.0
GitHub

About this Skill

Ideal for Codex-based AI Agents needing structured project workflow management using the .codex directory structure root is a skill that governs Codex project work for Hikaze-Model-Manager-2, maintaining stable sources of truth and tracking jobs and design details.

Features

Reads .codex/constitution/ for project principles and tech stack
Utilizes .codex/workflows/ for static workflow framework
References .codex/guidelines/ for implementation details and decisions
Tracks jobs and design details via .codex/jobs/activated
Maintains immutable sources of truth for project stability
Supports Hikaze-Model-Manager-2 for efficient project management

# Core Topics

HakureiHikaze HakureiHikaze
[0]
[0]
Updated: 3/6/2026

Quality Score

Top 5%
45
Excellent
Based on code quality & docs
Installation
SYS Universal Install (Auto-Detect)
Cursor IDE Windsurf IDE VS Code IDE
> npx killer-skills add HakureiHikaze/Hikaze-Model-Manager-2/root

Agent Capability Analysis

The root MCP Server by HakureiHikaze is an open-source Categories.community integration for Claude and other AI agents, enabling seamless task automation and capability expansion. Optimized for how to use root, root setup guide, what is root.

Ideal Agent Persona

Ideal for Codex-based AI Agents needing structured project workflow management using the .codex directory structure

Core Value

Empowers agents to maintain immutable sources of truth, track jobs, and manage design details by leveraging the .codex/constitution, .codex/workflows, and .codex/guidelines directories, ensuring consistency across Hikaze-Model-Manager-2 projects

Capabilities Granted for root MCP Server

Tracking project principles and tech stack in .codex/constitution
Managing static workflow frameworks in .codex/workflows
Implementing guidelines and decisions from .codex/guidelines
Activating and managing jobs from .codex/jobs/activated

! Prerequisites & Limits

  • Requires Hikaze-Model-Manager-2 compatibility
  • Limited to projects using the .codex directory structure
  • Immutable sources of truth must be predefined in the .codex directories
Project
SKILL.md
3.4 KB
.cursorrules
1.2 KB
package.json
240 B
Ready
UTF-8

# Tags

[No tags]
SKILL.md
Readonly

Root

Overview

Use this skill to govern Codex project work for Hikaze-Model-Manager-2 using the .codex directory structure, keeping immutable sources of truth stable while tracking jobs and design details.

Sources of truth and load order

  1. Read .codex/constitution/ for project principles, tech stack, and code standards.
  2. Read .codex/workflows/ for the static workflow framework.
  3. Read .codex/guidelines/ for implementation details and decisions.
  4. Read .codex/jobs/activated.md and the active job file to understand current work.

Hard rules

  • Do not edit .codex/constitution/ or .codex/workflows/ unless the user explicitly requests changes.
  • Propose suggestions for constitution/workflows only; do not modify without explicit user instruction.
  • Record new or changed design details in .codex/guidelines/ during design discussions.
  • Record designed but not implemented items in .codex/guidelines/not_implemented.md.
  • Keep all .codex content in Markdown.
  • When stating code or behavior facts, cite evidence (file path + line or command output); avoid assumptions.
  • Use a single file per job to carry phases and tasks.
  • The last task of every phase must be user manual verification.
  • Track branches as indented checklist items; complete same-indent branch items before returning to the parent flow.
  • Update .codex/jobs/activated.md whenever the active job changes.
  • If required .codex paths are missing, ask before creating them.

Job file format

Use jobs/<job>/<job>.md with Markdown checklists:

markdown
1# Job: <job-name> 2 3## Phase 1: <phase-name> 4- [ ] Task: <task> 5 - [ ] Branch: <subtask> 6 - [ ] Branch: <subtask> 7- [ ] Quality Gates 8 - [ ] Build passes (if applicable) 9 - [ ] Evidence recorded (file refs / command output) 10 - [ ] Docs updated (if needed) 11- [ ] User manual verification 12 13## Phase 2: <phase-name> 14- [ ] Task: <task> 15- [ ] Quality Gates 16 - [ ] Build passes (if applicable) 17 - [ ] Evidence recorded (file refs / command output) 18 - [ ] Docs updated (if needed) 19- [ ] User manual verification

Workflow usage

  • Treat workflows as the static guidance framework that governs each job and phase.
  • Record workflow branches in the job file as indented checklists.
  • Return to the parent flow after all items at the branch indent are complete.

Sub-skills

Use specialized skills to reduce context and improve precision when scope is narrow:

  • hikaze-frontend: Vue/TypeScript/CSS UI work in web/model_manager_frontend or web/custom_node_frontend.
  • hikaze-backend: Python nodes and server work in nodes/ or backend/.
  • hikaze-api: API contracts, request/response changes, and frontend-backend integration.

Conductor migration (only if requested)

Migrate prompts from custom_nodes/Hikaze-Model-Manager-2/conductor/ to .codex/ only when the user asks:

  • conductor/product.md -> .codex/constitution/product.md
  • conductor/tech-stack.md -> .codex/constitution/tech-stack.md
  • conductor/code_styleguides/* -> .codex/constitution/code_styleguides/*
  • conductor/workflow.md -> .codex/workflows/default.md
  • conductor/product-guidelines.md -> .codex/guidelines/product-guidelines.md
  • conductor/archive/* -> .codex/jobs/archive/*
  • Do not migrate conductor/discrepancies/* unless the user explicitly requests it.

Language policy

Accept any input language. Choose output language by accuracy first, then token efficiency.

Related Skills

Looking for an alternative to root or building a Categories.community AI Agent? Explore these related open-source MCP Servers.

View All

widget-generator

Logo of f
f

widget-generator is an open-source AI agent skill for creating widget plugins that are injected into prompt feeds on prompts.chat. It supports two rendering modes: standard prompt widgets using default PromptCard styling and custom render widgets built as full React components.

149.6k
0
Design

chat-sdk

Logo of lobehub
lobehub

chat-sdk is a unified TypeScript SDK for building chat bots across multiple platforms, providing a single interface for deploying bot logic.

73.0k
0
Communication

zustand

Logo of lobehub
lobehub

The ultimate space for work and life — to find, build, and collaborate with agent teammates that grow with you. We are taking agent harness to the next level — enabling multi-agent collaboration, effortless agent team design, and introducing agents as the unit of work interaction.

72.8k
0
Communication

data-fetching

Logo of lobehub
lobehub

The ultimate space for work and life — to find, build, and collaborate with agent teammates that grow with you. We are taking agent harness to the next level — enabling multi-agent collaboration, effortless agent team design, and introducing agents as the unit of work interaction.

72.8k
0
Communication