Turn one folder into your entire work base. A Claude Code plugin for operators and founders.
Run /setup once and Command Centre scaffolds an owned, reshapeable workspace: a contextualized CLAUDE.md that teaches Claude who you are, eleven purpose-built folders, and eleven forkable skills for your daily rhythm (capture, log, brief, draft, prep, decide, review). The scaffolding is universal; the contents are yours.
It is not a SaaS, a database, or a dashboard. It is a folder on your machine that Claude knows how to run.
You run a function largely on your own: growth, brand, ops, product, chief-of-staff, founder. You live in Claude Code and you're tired of re-explaining your context every session. You want one place that remembers, and tools that turn your daily motions into durable records instead of lost chat history.
If that's you, this is for you. If you want a team project-management app, it isn't.
You need Claude Code installed and working. Run these in any session, in this order.
1. Add the marketplace. Registers this repo so Claude Code can find the plugin:
/plugin marketplace add vanakrish/command-center
2. Reload so Claude Code picks up the new marketplace:
/reload-plugins
3. Install the plugin:
/plugin install command-center@command-center
When you install, Claude Code asks which scope to install at: user, project, or local. This choice matters, so pick deliberately:
- Project (recommended). Records the plugin in
.claude/settings.json, which is committed to git, so Command Centre travels with the repo to anyone who clones it. The right default for a workspace you want versioned and portable. - Local. Records it in
.claude/settings.local.json(gitignored), so it stays private to you and scoped to this one project. Choose this if you'd rather not commit the skills but still want them tied to this folder. It's also the easiest setup for editing the skill files as you go. - User (not recommended). Loads the skills globally, in every folder you ever open. Command Centre is built to run a single project or folder, so there's no reason to make it available everywhere.
No separate skill install. The eleven skills register with the plugin.
Run /setup before anything else. It's the one command that turns the plugin into your workspace:
/setup
It runs a short interview (~12 questions: who you are, your remit, key people, voice rules, what tools you have wired up), then writes three things into that folder:
CLAUDE.md: your operating context, so every future session already knows you.- Eleven folders, each with a README explaining what lives there.
.command-center/context.json: a machine-readable snapshot every skill reads from.
Then try /brain-dump testing my command centre to see the capture loop, and /daily-briefing tomorrow morning to start the rhythm.
| Skill | What it does |
|---|---|
/setup |
One-time: interview → builds your personalized workspace (folders + CLAUDE.md + .command-center/context.json) |
/brain-dump |
Capture a raw thought anytime; a bare invocation writes a session summary so you can safely /clear |
/daily-briefing |
Morning: calendar + carry-forwards + overnight comms + a top-3 focus → top of today's Daily Log |
/daily-log |
End of day: route Brain Dump into Decisions / To-dos / Ideas / Blockers + refresh CLAUDE.md workstreams |
/save-meeting |
Pull, structure, and save meetings, with a crash-safe catch-up loop |
/call-prep |
A pre-meeting brief from everything the workspace knows about a person or topic |
/draft-in-voice |
Write in your voice (or someone else's): three tone variants + an automatic humanizer pass |
/weekly-review |
Synthesize the week + opinionated next-week priorities |
/competitive-intel |
Scan your files + the web for competitor and market intel, tagged by source credibility |
/decision-log |
Capture a decision with options, rationale, and watchpoints |
/stakeholder-update |
Draft the recurring upward or outward update (a brief, not a status report) |
The skills are designed to compose, not just sit in a list:
- Morning.
/daily-briefingpulls your calendar, carry-forwards, and overnight comms to the top of today's Daily Log. - Throughout the day.
/brain-dumpappends raw, timestamped thoughts. Zero friction; no reformatting. - End of day.
/daily-logreads the Brain Dump and routes it into Decisions / To-dos / Ideas / Blockers, then refreshes the Active Workstreams block in yourCLAUDE.md.
Brain Dump is the raw tape; Daily Log is the processed view. Both stay. Over weeks, the folder becomes a searchable record of how you actually work.
/setup scaffolds these, each with a README explaining what lives there. The skills read and write here, so your context compounds in plain files you own, not a database you don't.
Output folders (skills write here):
| Folder | Holds |
|---|---|
Brain Dump/ |
Raw, timestamped captures throughout the day |
Daily Log/ |
Morning briefing + end-of-day routed sections + weekly reviews |
Meeting Transcripts/ |
Saved, summarized meetings |
Call Prep/ |
Pre-meeting briefs |
Drafts/ |
Content drafts in your voice |
Competitive Intel/ |
Competitor and market research |
Decisions/ |
One file per decision, with rationale and watchpoints |
Stakeholder Updates/ |
Recurring upward/outward updates |
Reference folders (you populate, skills read):
| Folder | Holds |
|---|---|
Brand/ |
Logos, fonts, colour tokens, brand guidelines |
Voice Samples/ |
Past writing that represents your voice, fuel for /draft-in-voice |
Working folder:
| Folder | Holds |
|---|---|
Scratch Pad/ |
Anything ephemeral, no conventions |
Several skills get materially richer when you've wired up tools: a calendar (/daily-briefing), a meeting-transcription tool (/save-meeting), team chat, email, a web-search tool (/competitive-intel), or a memory/semantic-search layer (/call-prep). Connecting them is where the system goes from a tidy filing cabinet to something that actually pulls your day together, so it's worth doing.
You don't have to set this up in advance. /setup asks you which connectors you have and records your answers. You can add or change connectors later by re-running /setup or editing .command-center/context.json.
And if you have none yet, everything still works. Skills that depend on a connector degrade gracefully: they fall back to your local files and tell you what they couldn't reach, rather than inventing it.
Every skill is just a SKILL.md file in this repo, yours to change. To customize one, copy its folder from this repo into your project's own skills folder and edit it there:
mkdir -p .claude/skills
# grab the skill from your clone of this repo (or download it from GitHub)
cp -R path/to/command-center/skills/daily-briefing .claude/skills/daily-briefing
A skill in your project's .claude/skills/ takes precedence over the plugin's copy, so your edited version wins. The scaffolding is universal; the bricks are yours.
- Install at project or local scope, not user. This is a per-project workspace, not a global tool. Project commits it to the repo so a team shares it; local keeps it private and easy to edit. User scope loads it everywhere for no good reason, so skip it.
- Run
/reload-pluginsafter adding the marketplace and after any update; it isn't always automatic. /setupis safe to re-run for a context refresh. It overwritesCLAUDE.mdandcontext.jsonbut preserves your existing folders and their contents.- Updating:
/plugin marketplace update command-center, then/reload-plugins. Your workspace (CLAUDE.md, folders, context.json) is untouched; only the skills update. Any skill you forked into.claude/skills/stays your version.
- Claude Code (recent version; the plugin uses the standard marketplace install).
- Works with any Claude model; Opus is the recommended default for this kind of knowledge work (
/model→ opus).
MIT © Krishna Veera Vanamali. See LICENSE.
Krishna Veera Vanamali · GitHub · whyvanamali.com