ai-agent-memory: Skill for Claude Code

New Engagement

When the user says "new engagement", "scaffold engagement", or invokes a slash command like /new-engagement, scaffold one or more client engagement projects from the resolved project template.

Resolve required roots

Before scaffolding:

1. Resolve WIKI_ROOT using memory/schema.md, then schema.md, then the memory-wiki path declared in AGENT.md.
2. Resolve TEMPLATE_ROOT by checking for project-template/ beside WIKI_ROOT, then the path declared in AGENT.md.
3. If the template is absent but network access is available, the agent may clone https://github.com/ozgurkarahan/ai-agent-memory.git into a temporary folder and use its project-template/ directory.
4. If either root remains unresolved, report the missing root and stop. Do not guess a personal path.
5. Remove any temporary clone after scaffolding.

All wiki paths below are relative to WIKI_ROOT.

Step 1: Parse the request

Extract from the user input:

• Client name (e.g., "Acme", "Contoso", "Contoso")
• Topics — one or more engagement topics, each with a format
• Projects root — use an explicit destination from the request or AGENT.md; otherwise ask where to create the client workspace

Expected input format: <ClientName> — <topic1> (format), <topic2> (format)

Formats: presentation, workshop, demo, presentation + demo, or combinations.

If the input is ambiguous or missing details, ask clarifying questions:

• What is the target audience? (e.g., technical leadership, developers, executives)
• What format? (presentation, workshop, demo, or combination)
• What are the key objectives?
• Any specific technologies or topics to cover?

Step 2: Scaffold each project

For each topic, do the following.

2a. Create project directory

PROJECT_DIR={PROJECTS_ROOT}/{ClientName}/10-projects/{project-slug}
mkdir -p "$PROJECT_DIR"

Convert the topic to a kebab-case slug (e.g., "Agent Framework Engagement" → agent-framework-engagement).