focus

Synchronize Spellbook primitives from a manifest across Claude Code, Codex, and Pi.

13|3|Updated Mar 2, 2026
One-click install
npx skills add https://github.com/phrazzld/spellbook --skill focus-phrazzld
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: focus
Source: https://github.com/phrazzld/spellbook/tree/main/skills/focus
Command: npx skills add https://github.com/phrazzld/spellbook --skill focus-phrazzld

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Focus configures the active set of Spellbook primitives for the current project by reading the .spellbook.yaml manifest, pulling skills and agents from GitHub, and coordinating harness-specific setup across Claude Code, Codex, and Pi.

Core Features & Use Cases

  • Resolve and synchronize project-local domain primitives from a manifest across all harness targets.
  • Nuke-and-rebuild managed primitives on every run to ensure project consistency.
  • Provide a guided workflow for focusing on relevant primitives, syncing changes, and managing the project-specific skill set.

Quick Start

Analyze the project with /focus init to generate a .spellbook.yaml manifest.

Frequently Asked Questions about focus

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I manage project-specific skills and agents across Claude Code, Codex, and Pi?

To manage project-specific skills across Claude Code, Codex, and Pi, you can use a manifest file to coordinate active primitives and apply harness-agnostic loading logic for consistent deployments across all targets.

How do I sync Spellbook primitives from a GitHub repository to my local project?

You can sync Spellbook primitives by reading a manifest configuration that pulls required skills and agents directly from GitHub, applying nuke-and-rebuild logic to ensure your local project remains consistent with the source.

What is a nuke-and-rebuild strategy for managing development environment primitives?

A nuke-and-rebuild strategy for development primitives removes and completely reconstructs managed skills on every run, ensuring your project configuration stays strictly consistent with the manifest without leftover legacy files.

Can I use a single manifest to deploy domain primitives across multiple AI coding harnesses?

Yes, you can use a single manifest to deploy domain primitives across multiple AI coding harnesses by applying harness-agnostic loading logic that coordinates setup across Claude Code, Codex, and Pi simultaneously.

How do I initialize a project manifest to orchestrate relevant coding skills?

To initialize a project manifest for orchestrating coding skills, analyze your project workspace to automatically generate a configuration file, which then guides the workflow for focusing on and syncing relevant primitives.

Why do my project skills keep desyncing across different AI coding tools?

Project skills desync across different AI coding tools when configurations are managed independently per environment; resolving this requires reading a central manifest and applying nuke-and-rebuild logic to enforce consistency across all harnesses.