agent-flow-init-check

Check for agent-flow initialization files before running commands.

2|1|Updated Apr 9, 2026
One-click install
npx skills add https://github.com/timgranlundmarsden/claude-agent-flow --skill agent-flow-init-check
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-flow-init-check
Source: https://github.com/timgranlundmarsden/claude-agent-flow/tree/main/skills/agent-flow-init-check
Command: npx skills add https://github.com/timgranlundmarsden/claude-agent-flow --skill agent-flow-init-check

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents agent-flow commands from running in a repository that has not been initialized, avoiding confusing failures and making the required setup state explicit.

Core Features & Use Cases

  • Initialization Guard: Verifies that the repository has either the installed sync state file or the master source manifest before any command continues.
  • Source vs Installed Repo Detection: Recognizes the master repository through the publish manifest and skips the check silently when that file is present.
  • Fail-Fast Enforcement: Stops execution immediately and tells the user to run /install when the repository is not ready.
  • Use Case: A user tries to invoke an agent-flow command in a fresh clone, and the Skill blocks the action until initialization is completed.

Quick Start

Ask the agent-flow guard to verify this repository is initialized and stop with the /install message if neither required state file is present.

Frequently Asked Questions about agent-flow-init-check

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

FAQPage Schema
How do I prevent agent-flow commands from running before repository initialization in Claude Code?

Repository initialization gating blocks agent-flow commands by checking for .claude-agent-flow/publish-plugin-manifest.yml or sync-state.json, halting execution with a /install instruction if neither file exists.

Why do my agent-flow commands fail immediately in a fresh clone?

Commands fail in a fresh clone because the repository initialization check cannot find the required sync-state.json or publish-plugin-manifest.yml file, triggering a fail-fast halt until you run /install.

How does command gating differentiate between installed projects and the master source repository?

Command gating identifies the master source repository by detecting the publish-plugin-manifest.yml file and skips the initialization check silently, while installed projects require the sync-state.json file.

What repository files are required to bypass workflow control initialization checks?

To bypass initialization checks, a repository must contain either the .claude-agent-flow/sync-state.json file for installed projects or the .claude-agent-flow/publish-plugin-manifest.yml file for the master source.

When should I run the install command for Claude Agent Flow repositories?

You should run the install command when the initialization guard blocks command execution, which happens when neither the sync-state.json nor the publish-plugin-manifest.yml file is present in the .claude-agent-flow directory.

Can I execute agent-flow commands without setting up the .claude-agent-flow directory?

No, executing commands without setup is blocked by the initialization guard, which immediately halts workflow control and requires the .claude-agent-flow directory to contain valid state files before proceeding.