# Brain Setup > > - **Type:** Skill - **Install:** `agentstack add skill-cerebrocybersolutions-libro-brain-setup` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [cerebrocybersolutions](https://agentstack.voostack.com/s/cerebrocybersolutions) - **Installs:** 0 - **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml) - **Latest version:** 0.1.0 - **License:** Apache-2.0 - **Upstream author:** [cerebrocybersolutions](https://github.com/cerebrocybersolutions) - **Source:** https://github.com/cerebrocybersolutions/libro/tree/main/skills/brain-setup - **Website:** https://cerebrocybersolutions.com/libro ## Install ```sh agentstack add skill-cerebrocybersolutions-libro-brain-setup ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # Brain Setup Guided first-time setup that builds the Brain folder system from scratch. Prompts for operator info and departments, scaffolds the full structure on disk, seeds a working DASHBOARD.md + awareness.md + workspace CLAUDE.md + per-dept CLAUDE.md files, and hands off with clear instructions for connecting the session rituals (sessionstart + sessionend). **Time to complete:** ~7 minutes. --- ## ⚠️ FIRST-TIME SKILL — No Brain Check This skill *creates* the Brain. There is nothing to check yet. Skip any Brain Check. Begin directly with Stage 1. --- ## STAGE ROUTING | Where the operator is | Stage | Reference file | |---|---|---| | Just started — no info collected yet | Stage 1: Intake | `references/stage1_intake.md` | | Profile confirmed, ready to build folders + docs | Stage 2: Scaffold | `references/stage2_scaffold.md` | | Folders built, ready to verify + connect session rituals | Stage 3: Connect | `references/stage3_connect.md` | | Running full setup end-to-end | All stages in sequence | Read all three reference files | --- ## DEFAULT DEPARTMENTS (example pattern) If the operator has no custom departments in mind, offer these as a starting point. Folder names are lowercase kebab-case; display names are friendly for docs. | Code | Folder (kebab-case) | Display name | Purpose | |---|---|---|---| | ops | `ops/` | Operations | Company infrastructure, SOPs, skills, dashboards | | sales | `sales/` | Sales | Pipeline, outreach, proposal tracking | | delivery | `delivery/` | Delivery | Client work, projects, service execution | | products | `products/` | Products | Product builds, packaging, pricing | | content | `content/` | Content | Marketing, social, content ops | Always ask before applying. Accept any list the operator provides. Custom departments are welcome — accept any list, auto-convert names to kebab-case, confirm with operator before building. --- ## WORKSPACE PATHS All paths are collected from the operator at runtime in Stage 1. Nothing is hardcoded — this skill works for any operator on any machine. After Stage 1, the pattern is: ``` {workspace_root}/ ← top-level project folder {workspace_root}/CLAUDE.md ← workspace-root front door (created in Stage 2) {workspace_root}/mission-control/ ← Master Brain (Chief of Staff) {workspace_root}/mission-control/DASHBOARD.md {workspace_root}/mission-control/awareness.md {workspace_root}/mission-control/sessions/ {workspace_root}/mission-control/decisions/ {workspace_root}/mission-control/skills/ ← skill source-of-truth layer {workspace_root}/mission-control/state/ ← live state {workspace_root}/{dept-folder}/ ← one per department {workspace_root}/{dept-folder}/CLAUDE.md ← dept-scoped front door {workspace_root}/{dept-folder}/brain/ {workspace_root}/{dept-folder}/brain/README.md {workspace_root}/{dept-folder}/brain/sessions/ {workspace_root}/{dept-folder}/brain/decisions/ {workspace_root}/{dept-folder}/brain/decisions/decisions.md {workspace_root}/{dept-folder}/brain/pipeline/ ``` --- ## SURFACE PRIMER (seeded into workspace CLAUDE.md) The workspace CLAUDE.md seeds this primer so every new session knows where tasks belong: - **Claude Code** = on-metal executor. Runs in the operator's terminal with Mac/Linux FS, local models, and git. Surface for drafting, planning, synthesis, file ops, repo work, local model runs, anything that touches process-state on the machine. - **Operator** = relationship + decision layer. Human judgment, vendor/customer touches, sign-offs. Routing rule: Everything goes through Claude Code unless the Operator needs to step in for a relationship or hard decision. --- ## KEY RULES 1. **Confirm before building.** Present the full profile and folder plan to the operator and wait for a 'yes' before running any commands. A wrong root path is painful to undo. 2. **Never overwrite existing content.** Before creating any folder or file, check if it exists. If a Brain is partially built, surface what's there and ask how to proceed — never silently skip or clobber existing files. 3. **Folders use lowercase kebab-case.** `ops/`, not `Ops/`. `content-creation/`, not `ContentCreation/`. Custom dept names get auto-converted. 4. **Scaffold the docs, don't leave them empty.** Every folder the skill creates gets a starter doc (CLAUDE.md, README.md, DASHBOARD.md, awareness.md, decisions.md). A Brain with empty folders is indistinguishable from a broken Brain. 5. **Stop at instructions for session rituals — do not edit them.** Give the operator the Brain root path + explicit what-to-change-where for `sessionend` AND `sessionstart`. Don't reach into another skill's files. 6. **Verify before signing off.** After scaffolding, list every folder + file that was created and confirm each exists. A setup that silently fails is worse than one that errors loudly. 7. **Never seed operator-specific data.** This skill ships to external operators. Do not embed any operator-specific identifiers into the scaffolded files. The operator supplies their own profile via Stage 1 intake — the skill scaffolds the structure, the operator owns the content. --- ## Scope Contract | Dimension | Scope | |-----------|-------| | Read paths | Operator input collected via Stage 1 prompts (name, company, workspace root, department roster) | | Write paths | New `{workspace_root}/`, `{workspace_root}/mission-control/**`, `{workspace_root}/{dept}/brain/**`, and starter docs — written once during Stage 2 scaffold | | MCP / tool surface | Python stdlib (`os`, `pathlib`) for folder creation; file I/O for starter docs | | Network egress | None | | Surface | Claude Code | | Credentials | None | | Escalation trigger | If target root has existing content → **HALT and surface** per Rule 2; require operator confirmation or new path before any write | ## References (auto-wikilinks) - [[DASHBOARD]] - [[awareness]] - [[decisions]] - [[stage1_intake]] - [[stage2_scaffold]] - [[stage3_connect]] ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [cerebrocybersolutions](https://github.com/cerebrocybersolutions) - **Source:** [cerebrocybersolutions/libro](https://github.com/cerebrocybersolutions/libro) - **License:** Apache-2.0 - **Homepage:** https://cerebrocybersolutions.com/libro Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** yes - **Shell / process execution:** no - **Environment & secrets:** no - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/skill-cerebrocybersolutions-libro-brain-setup - Seller: https://agentstack.voostack.com/s/cerebrocybersolutions - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.