start

name	start
description	Initialize tasks and memory and open the dashboard — a light first-run wedge without the full profile interview. Use when setting up for the first time, bootstrapping working memory from an existing task list, or decoding shorthand in todos.

Start command

If you see unfamiliar placeholders or need to check which tools are connected, see CONNECTORS.md.

Initialize the task and memory systems, then open the unified dashboard. This is the fast path — for identity, voice, VIP tiers, and email/calendar policies, run /assistant:setup afterward.

Instructions

1. Check what exists

Resolve the working folder per rules/paths.md (default ~/MyAssistant). Check for:

• TASKS.md — task list
• AGENTS.md — working memory hot cache
• CLAUDE.md — compat shim (@AGENTS.md)
• memory/ — deep memory directory
• dashboard.html — visual UI

2. Create what's missing

If TASKS.md doesn't exist: create it from skills/assistant-setup/assets/TASKS.template.md (sections: To do, In progress, In review, Done, Cancelled, Blocked).

If dashboard.html doesn't exist: copy from ${CLAUDE_PLUGIN_ROOT}/skills/dashboard.html to the working folder.

If AGENTS.md, CLAUDE.md, and memory/ don't exist: scaffold from setup assets (AGENTS.template.md, CLAUDE.template.md, glossary.template.md, memory/people/, memory/projects/, memory/context/). Then continue to memory bootstrap (step 5).

Do not write config/profile.md or policies in this command — offer /assistant:setup when the user wants inbox/calendar sharpness.

3. Open the dashboard

Do not use open or xdg-open — in Cowork, the agent runs in a VM and shell open commands won't reach the user's browser. Tell the user:

Dashboard is ready at {assistantPath}/dashboard.html. Open it from your file browser (Chrome or Edge) and select this folder.

4. Orient the user

If everything was already initialized:

Dashboard ready. Tasks and memory are loaded.
- /assistant:update — sync tasks, memory gaps, and follow-ups
- /assistant:update --all — deep scan of connected email, calendar, chat, and tasks
- /assistant:setup — full profile, voice, and policies (recommended when you connect mail)

If memory hasn't been bootstrapped yet, continue to step 5.

5. Bootstrap memory (first run only)

Only when AGENTS.md and memory/ are empty or freshly scaffolded.

The best source of workplace language is the user's actual task list. Real tasks = real shorthand.

Ask the user:

Where do you keep your todos? This could be:
- A local file (e.g. TASKS.md, todo.txt)
- An app (Asana, Linear, Jira, Notion, Todoist, GitHub Issues)
- A notes file or paste

I'll use your tasks to learn your workplace shorthand.

Once you have the task list:

For each task item, analyze shorthand — names, acronyms, project codenames, internal terms. Decode interactively using the checklist in skills/memory-management/SKILL.md (✓ / ? markers). Ask only about terms not already in memory.

6. Optional comprehensive scan

After task decoding, offer:

Want a deeper scan of chat, email, and docs for people and projects?
Run /assistant:update --all when connectors are ready — or skip for now.

If they choose to scan now and connectors are available, follow the comprehensive templates in commands/update.md (--all section). Group memory proposals by confidence (see memory-management skill).

7. Write memory files

From everything gathered, populate:

• AGENTS.md — People, Terms, Projects, Preferences tables (~50–80 lines). Link to config/profile.md when a profile exists.
• memory/glossary.md — full decoder ring
• memory/people/{slug}.md, memory/projects/{slug}.md, memory/context/ as needed

Present memory proposals grouped by confidence:

• Ready to add (high confidence) — offer to add directly
• Needs clarification — ask the user
• Low frequency / unclear — note for later

8. Report results

Start complete:
- Tasks: TASKS.md ({N} items)
- Memory: {N} people, {N} terms, {N} projects
- Dashboard: open in browser

Next: /assistant:update to stay current · /assistant:setup for profile and voice

Notes

• If memory is already populated, this command scaffolds missing files and points at the dashboard.
• Nicknames are critical — always capture how people are actually referred to.
• If a source isn't available, skip it and note the gap.
• Memory grows organically through conversation after bootstrap.