forgetful-project-init

name	forgetful-project-init
description	Create or initialize Forgetful project records for the current repository by matching the current working directory to an existing Forgetful project, asking before creating a missing project, and maintaining generic device/location metadata. Use when the user asks to initialize, create, register, or set up Forgetful project context.

Forgetful Project Init

Use this skill when the user wants to create, register, or initialize a Forgetful project for the current repository.

This skill is for project existence and setup only. Do not fetch recent memories as part of this workflow; use $forgetful-project-load when the user wants recent project context.

Workflow

1. Determine the current working directory from session context.
2. Discover project candidates:

execute_forgetful_tool("list_projects", {})

3. Match the current directory to a Forgetful project. Prefer repository metadata from the git remote when available, then project name or project notes. Do not guess a project ID.
4. If a project matches, briefly mention the matched project and ID. Do not load memories.
5. If no project matches, ask the user whether they want to create a Forgetful project for this directory. Do not create it without confirmation.
6. When a project exists or the user confirms creation, maintain a device/location note:
   • Discover the current device from runtime context, such as hostname or environment details available in the session.
   • Search for an existing device entity before creating one.
   • Link the device entity to the project.
   • Create or update one project-scoped memory noting this project's local path on that device.

Project Creation

If the user confirms project creation, use:

execute_forgetful_tool("create_project", {
  "name": "<project name>",
  "description": "<short purpose/scope>",
  "project_type": "development",
  "repo_name": "<owner/repo when known>"
})

Choose a conservative project type from the available Forgetful schema. Use development unless the project is clearly documentation, infrastructure, personal, open-source, or another supported type.

Privacy And Distribution

This skill is distributed publicly. Do not hardcode or assume private names, hostnames, local paths, repository owners, or user-specific conventions. Use only runtime-discovered values and user-confirmed information.

Do not save secrets or credentials in project, entity, or memory notes.