菜单
Create and manage Initra-compliant Agent Skills with proper progressive disclosure structure (Level 1-3). Use when you need to automate specialized workflows, encapsulate framework expertise, or extend agent capabilities through reusable procedural knowledge.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Orchestrate specialized agent teams (Architect, Coder, Tester, Reviewer) using high-fidelity communication protocols. Use when a task requires cross-disciplinary coordination, complex architectural changes, or multi-agent parallel workflows.
Guide for creating high-fidelity Venture Blueprints. Use when adding new templates or project types to the Initra catalog.
Helps with debugging and expanding the Initra deterministic prompt engine. Use when modifying src/lib/engine/ logic, templates, or package/service libraries.
| name | skill-builder |
| description | Create and manage Initra-compliant Agent Skills with proper progressive disclosure structure (Level 1-3). Use when you need to automate specialized workflows, encapsulate framework expertise, or extend agent capabilities through reusable procedural knowledge. |
This skill provides the framework for creating production-ready Agent Skills within the Initra ecosystem. It ensures all new skills follow the 3-level progressive disclosure standard for maximum token efficiency.
---.Define the domain of the new skill.
Example: supabase-integration, pino-logging-standards, nextjs-proxy-management.
Create the top-level structure:
mkdir -p .agents/skills/[skill-name]/docs
mkdir -p .agents/skills/[skill-name]/examples
mkdir -p .agents/skills/[skill-name]/scripts
Follow the template provided in AGENTS.md. Ensure the description uses high-density trigger words to enable autonomous matching.
name and description.AGENTS.md catalog.