# Install Architect Playbook Globally > Copy every architect-playbook skill into ~/.claude/skills/ so the slash commands are available in every Claude Code session on this machine. - **Type:** Skill - **Install:** `agentstack add skill-bensheridanedwards-architectplaybook-install-architect-playbook-globally` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [BenSheridanEdwards](https://agentstack.voostack.com/s/bensheridanedwards) - **Installs:** 0 - **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills) - **Latest version:** 0.1.0 - **License:** MIT - **Upstream author:** [BenSheridanEdwards](https://github.com/BenSheridanEdwards) - **Source:** https://github.com/BenSheridanEdwards/ArchitectPlaybook/tree/main/install-architect-playbook-globally ## Install ```sh agentstack add skill-bensheridanedwards-architectplaybook-install-architect-playbook-globally ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # /install-architect-playbook-globally Copy every skill folder from this architect-playbook repository into `~/.claude/skills/`. After running this, every architect-playbook slash command works in every Claude Code session on the machine, regardless of which project is open. This skill is the machine-wide counterpart to [`/install-architect-playbook-locally`](../install-architect-playbook-locally/SKILL.md). The mechanics are the same; only the destination differs. ## Usage ``` /install-architect-playbook-globally # install or update every skill into ~/.claude/skills/ /install-architect-playbook-globally --dry-run # print the plan without copying anything /install-architect-playbook-globally --force # overwrite destinations even if they appear newer /install-architect-playbook-globally --include=name # install only one named skill (repeatable) /install-architect-playbook-globally --exclude=name # skip a named skill (repeatable) ``` ## When to choose this over the local install Use this skill when: - You audit lots of different codebases and you do not want to re-install per project. - You want a single canonical version of every skill on disk. Prefer `/install-architect-playbook-locally` when: - You want different projects to pin different versions of the playbook. - You want the installed skills to live alongside the project in version control. ## What this skill does 1. **Resolves the playbook root** the same way `/install-architect-playbook-locally` does. 2. **Resolves the destination** as `$HOME/.claude/skills/`. Creates it if missing. 3. **Asks for confirmation** before overwriting any existing skill of the same name in `~/.claude/skills/`. Global installs affect every Claude Code session on the machine, so the bar for overwriting must be higher than for a project-local install. 4. **Enumerates source skills** (every sub-folder with a `SKILL.md` except `install-architect-playbook-locally` and `install-architect-playbook-globally`). 5. **Copies each skill folder** with `cp -R`. The destination becomes a complete, self-contained copy. 6. **Reports the result** with the same format as the local installer. ## Implementation steps ### Step 1 — Resolve the playbook root Same logic as `/install-architect-playbook-locally`. The skill file's parent of parent is the playbook root. ### Step 2 — Resolve destination ```bash DEST="$HOME/.claude/skills" mkdir -p "$DEST" ``` Refuse to proceed if `$HOME` is unset or empty. ### Step 3 — Enumerate skills Same as the local installer. Exclude: - `install-architect-playbook-locally` - `install-architect-playbook-globally` - Any folder whose name begins with `.` ### Step 4 — Confirmation gate for overwrites Before copying, list every destination that already exists: ``` The following skills already exist in ~/.claude/skills/ and will be overwritten: - security-audit - performance-audit - graphify (NOT MANAGED BY THIS PLAYBOOK — will not be touched) Continue? [y/N] ``` If the user does not answer `y`, abort. Do not overwrite skills the playbook does not own — for the architect-playbook this is everything except `graphify`. Detect ownership by checking whether the destination's `SKILL.md` `name` field matches a skill in the playbook source. Anything whose name is not in the source list is left alone. ### Step 5 — Copy each skill Same loop as the local installer, but writing to `$HOME/.claude/skills/`. ### Step 6 — Print the summary ``` installed: pre-audit-setup updated: security-audit installed (stub): performance-audit ... preserved: graphify (not managed by architect-playbook) 15 skills installed into ~/.claude/skills/ Open a new Claude Code chat (any project) to pick up the new slash commands. ``` ## Idempotency and safety rules - Never overwrite skills the playbook does not own. `graphify` is the canonical example. - Re-running with no flags is safe. - Refuses to write outside `$HOME/.claude/skills/`. - Refuses to delete any directory in the destination — only `cp -R` over the top of it. - `--dry-run` is honored. ## Recommended commit message This skill writes to your home directory, not your project, so there is nothing to commit. If you maintain `~/.claude` as a tracked dotfiles repository, use: ``` chore: refresh architect-playbook skills in ~/.claude/skills/ ``` ## What this skill explicitly does NOT do - Install graphify. - Modify any project's `.claude/settings.json`. - Run any audit. - Touch global Claude Code settings (`~/.claude/settings.json`). ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [BenSheridanEdwards](https://github.com/BenSheridanEdwards) - **Source:** [BenSheridanEdwards/ArchitectPlaybook](https://github.com/BenSheridanEdwards/ArchitectPlaybook) - **License:** MIT 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:** no - **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-bensheridanedwards-architectplaybook-install-architect-playbook-globally - Seller: https://agentstack.voostack.com/s/bensheridanedwards - 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%.